Who decides where a drug sits on the federal schedules, and what constrains that decision? Drug scheduling is the administrative process by which substances move onto, between, and off the five schedules of the Controlled Substances Act, and the answer to both halves of the question lies in a single unusual bargain struck in 1970. Congress divided the power between two agencies: the enforcement agency that initiates proceedings and writes the final orders, and the health agency whose scientific judgment can stop control cold but cannot compel it. The result is a system in which placement is never one office’s call, in which the evidentiary demands are written into the statute itself, and in which international treaty commitments and an emergency lane for imminent hazards define the outer edges of what the domestic machinery may do and may undo.

The placement of a substance carries consequences that reach far beyond a label. A schedule determines whether a drug may be prescribed at all, whether prescriptions may be refilled, what registration and security obligations manufacturers and researchers carry, and what criminal penalties attach to unauthorized distribution. Because so much turns on the outcome, the statute builds the decision the way it builds a trial: with named findings that must be made, an evidentiary checklist that must be worked through, a health agency whose negative verdict ends the matter, a hearing in which any affected interest may test the evidence, and a reviewing court that examines the record. This article walks that machinery in the order a proceeding actually travels. It opens with the statutory gate, the three findings that must be made for each schedule, and shows how the five schedules descend from prohibition to ordinary pharmacy practice. It then traces the three routes into a proceeding, the petitions any interested party may file, the health agency evaluation that follows, and the one-directional veto that evaluation carries. The eight factors the statute commands the agencies to consider come next, with the documented record of how each factor has been applied to real substances. From there the account turns to the two provisions that sit outside the ordinary machinery, temporary scheduling for imminent hazards and the treaty provision that limits how far control may be undone, and then to the most argued-about placement in the system, the Schedule I listing of cannabis, the circularity rescheduling petitioners identify in it, and the record of substances that have moved in both directions. The final section states the honest account: a system with genuine constraints that nonetheless moves downward far more slowly than it moves up, and explains why that tempo is a product of the statute’s design.

The Three Findings Behind Drug Scheduling

The scheduling architecture of the Controlled Substances Act rests on a single statutory gate. Section 201 of the findings, codified at 21 U.S.C. 812(b), commands that a drug or other substance may not be placed in any schedule unless the findings required for that schedule are made with respect to it. The language is prohibitory rather than directory: placement without the matching findings is forbidden, not merely discouraged. Two narrow exceptions cut through the rule. Control required by United States obligations under international treaties, conventions, or protocols in effect on the effective date of the statute passes without the findings, and so does control of an immediate precursor, on the logic that a chemical that exists only to become an already controlled substance needs no independent analysis of its own medical value or abuse risk. Everything else travels through the findings.

The findings draw their evidence from the eight-factor record compiled under 21 U.S.C. 811(c), and the statute assigns the scientific portion of that record to the Department of Health and Human Services. Under 21 U.S.C. 811(b), the Secretary’s scientific and medical recommendations bind the Attorney General as to scientific and medical matters, and a recommendation against control ends the proceeding. That binding relationship, examined in full later in this article, is the reason the findings are never the product of one agency alone. The medical status of a substance is decided by the health authorities, and the placement is the product of that binding medical judgment combined with the enforcement judgment that follows it.

The findings themselves reduce to three inquiries, restated with small but load-bearing variations for each schedule. The first inquiry asks about potential for abuse. Every schedule except the first frames this comparatively: abuse potential less than the substances in schedules I and II for schedule III, low potential relative to schedule III for schedule IV, low potential relative to schedule IV for schedule V. Only schedules I and II share the unqualified finding of high potential for abuse, a pairing that carries a consequence worth stating plainly. On abuse risk alone, the statute treats schedule I and schedule II substances as identical. What separates a schedule I narcotic from a schedule II medicine is never how much abuse risk the substance carries. It is everything else.

The second inquiry asks whether the substance has a currently accepted medical use in treatment in the United States, the statute’s own phrase. This is the hinge on which the entire system turns. Schedule I demands a negative finding: no currently accepted medical use. Schedule III, schedule IV, and schedule V all demand a positive one: currently accepted medical use in treatment in the United States. Schedule II occupies a middle position with a built-in alternative. Its medical use finding can be satisfied either by a currently accepted medical use in treatment in the United States or by a currently accepted medical use with severe restrictions, a formulation that accommodates substances whose legitimate therapeutic role survives only inside tightly controlled distribution or administration protocols. The distinction between accepted medical use and accepted medical use with severe restrictions has no counterpart anywhere else in the findings.

The third inquiry splits in two directions depending on the schedule. For schedule I it asks about safety under supervision: there must be a lack of accepted safety for use of the drug or other substance under medical supervision. For the remaining schedules it asks about dependence, calibrated on a descending scale. Schedule II requires a finding that abuse may lead to severe psychological or physical dependence. Schedule III lowers the finding to moderate or low physical dependence or high psychological dependence. Schedule IV lowers it again to limited physical dependence or psychological dependence relative to the substances in schedule III. Schedule V reaches the floor: limited physical dependence or psychological dependence relative to the substances in schedule IV. The dependence inquiry thus measures not only how much harm abuse causes but how much harm it causes compared with the substances already sorted into the neighboring schedules.

The decisive legal point is that these findings operate conjunctively, and the conjunctive force is strongest at the top of the system. Placement in schedule I requires all three findings at once: high potential for abuse and no currently accepted medical use in treatment in the United States and a lack of accepted safety for use under medical supervision. A substance that fails any one of the three cannot lawfully sit in schedule I. High abuse potential alone does not qualify a substance for the most restrictive schedule, because the schedule I finding set treats abuse risk as only one of three necessary conditions. Nor does the absence of medical use qualify a substance for schedule I by itself, since schedule I also demands the separate safety finding. The structure deliberately makes schedule I difficult to occupy in full, because schedule I carries the most severe regulatory consequences: no lawful prescribing, no pharmacy dispensing, and acquisition permitted only for federally sanctioned research.

The medical use finding has generated more scheduling litigation and administrative proceedings than the other two combined, which is expected since medical acceptance is a judgment about evolving science rather than a measurement. When the Department of Health and Human Services concluded in 2023 that marijuana has a currently accepted medical use in treatment in the United States, that conclusion struck directly at the second prong of the schedule I test. The Department of Justice published a notice of proposed rulemaking on May 21, 2024, to transfer marijuana from schedule I to schedule III, reasoning that marijuana’s potential for abuse is less than the substances in schedules I and II and that its abuse may lead to moderate or low physical dependence or high psychological dependence, which is the schedule III dependence formulation nearly verbatim. The proposed transfer illustrates the conjunctive logic in motion. Marijuana’s long occupancy of schedule I depended on all three findings holding simultaneously; once the medical use finding was reconsidered, the schedule I placement lost one of its three legs and the proposed destination became schedule III rather than schedule II, because the proposed findings matched the schedule III descriptors rather than the schedule II ones. The statute requires such a transfer to proceed through formal rulemaking on the record after an opportunity for a hearing, so the proposed rule opened a proceeding rather than changing the law by itself.

The Five Schedules, From Most Restrictive to Least

With the three findings in place, the five schedules read as a single descending argument. Schedules I and II share the high abuse potential finding and diverge entirely on medical use and safety. Schedules II through V share the positive medical use finding and diverge on a steadily diminishing assessment of abuse potential and dependence severity, each schedule measuring itself against the one above it. The relative framing means that every placement below schedule I is inherently comparative: a substance belongs in schedule III not because its abuse potential sits at some absolute level but because it sits below the schedule I and II population and above the schedule IV one. That relativity is why rescheduling proceedings so often move substances between adjacent tiers rather than across the whole ladder, and why the findings function, as section 812(b) requires, as a gate that must be opened anew for each substance and each schedule, with all three findings made, every time, on the record evidence compiled under the eight factors of section 811(c).

What makes Schedule I different from Schedule II?

Both schedules demand a finding of high abuse potential, so abuse potential cannot separate them. The difference lies in the other two findings: Schedule I requires no accepted medical use and a lack of accepted safety under supervision, while Schedule II requires accepted medical use and severe dependence liability. Medical value, not danger, draws the line between prohibition and restriction.

The original schedules written directly into the statute in 1970 populated schedule I with the substances Congress judged to meet all three findings at the time: heroin, lysergic acid diethylamide, methaqualone, peyote, and marijuana among them. Heroin remains the archetypal schedule I narcotic, the example against which the schedule is most often explained, because it combines all three required characteristics in the clearest form. LSD sits beside it as the archetypal schedule I hallucinogen. Methaqualone, a sedative hypnotic once sold under brand names and later withdrawn from the market, shows that schedule I status does not require a substance to be a street drug in the narcotic sense; it requires only that the three findings be met. These placements were made by Congress in the statute itself, so they did not pass through the administrative findings process that later additions must survive.

Schedule II holds substances that share schedule I’s high abuse potential but possess a currently accepted medical use, which is precisely why the two schedules can carry identical abuse findings and radically different legal treatment. Cocaine, methamphetamine, morphine, hydromorphone, methadone, oxycodone, fentanyl, and the amphetamine and methylphenidate products sold as Adderall, Dexedrine, and Ritalin all sit in schedule II. Each is a medicine with real therapeutic indications; each carries high abuse potential; abuse of each may lead to severe psychological or physical dependence. The schedule II population is the statute’s acknowledgment that dangerousness and medical value can coexist in the same molecule, and the regulatory response is restriction rather than prohibition. Schedule II prescriptions cannot be refilled, for example, a rule that follows directly from the dependence finding and that distinguishes the schedule from everything below it.

The most instructive schedule II placement decision of the past several decades involved hydrocodone combination products such as Vicodin, Lortab, and Norco. The combination products had sat in schedule III for decades on the assumption that the combination form carried a lesser likelihood of abuse, while hydrocodone itself sat in schedule II. By the early 2010s, diversion and misuse data had undercut that assumption, and after a scientific and medical evaluation the agency proposed and then finalized the move to schedule II, effective October 6, 2014. The medical use finding stayed positive; the abuse potential finding was revised upward to the schedule II formulation, and the dependence finding followed it. The full procedural record of that transfer, including the 1999 petition and the 2013 advisory committee vote that preceded it, is set out later in this article alongside the other documented movements between schedules.

Schedule III marks the first point in the descending sequence where abuse potential is framed comparatively rather than in absolute terms: the substance must have a potential for abuse less than the substances in schedules I and II. The medical use finding becomes the straightforward positive version, currently accepted medical use in treatment in the United States, and the dependence finding drops to moderate or low physical dependence or high psychological dependence. The products containing less than ninety milligrams of codeine per dosage unit, exemplified by Tylenol with codeine, were among the original schedule III listings and remain its most familiar members: narcotic enough to require control, dilute enough in combination form to fail the schedule II dependence test. Ketamine joined schedule III through a final rule published in July 1999, after the Drug Enforcement Administration found that the anesthetic’s growing nonmedical use met the schedule III abuse finding while its established role in human and veterinary anesthesia preserved the medical use finding. The dissociative anesthetic’s trajectory from unscheduled medicine to schedule III control shows the same mechanism as the hydrocodone story in the opposite direction along the abuse gradient, with the medical use finding acting as the anchor that kept the substance out of schedule I.

Anabolic steroids reached schedule III by statute rather than by rulemaking. The Anabolic Steroids Control Act of 1990, enacted as Title XIX of the Crime Control Act of 1990 and effective February 27, 1991, placed anabolic steroids into schedule III of the Controlled Substances Act and defined the class as drugs or hormonal substances chemically and pharmacologically related to testosterone, excluding estrogens, progestins, and corticosteroids, that promote muscle growth. The 1990 Act also addressed human growth hormone and established criminal penalties for nonmedical use and distribution. Because the scheduling was legislative, it bypassed the administrative findings process, though Congress plainly had the schedule III findings in mind: legitimate therapeutic uses for wasting and hormonal conditions alongside a pattern of nonmedical use for performance enhancement with dependence liability below the schedule II threshold. A later statute, the Anabolic Steroid Control Act of 2004, added steroid precursors to the same schedule, extending the original legislative judgment rather than reopening it.

Schedule IV descends one more step on both gradients that define the system: low potential for abuse relative to the substances in schedule III, and abuse that may lead only to limited physical dependence or psychological dependence relative to the substances in schedule III, while the medical use finding remains the positive one. The schedule’s population is dominated by the medicines prescribed most widely for anxiety, insomnia, and moderate pain. Alprazolam, sold as Xanax; diazepam, sold as Valium; lorazepam, sold as Ativan; clonazepam, sold as Klonopin; and zolpidem, sold as Ambien, all sit in schedule IV. The carisoprodol product Soma sits there as well. These are substances whose therapeutic roles are broad and long established, whose abuse patterns are real but comparatively modest, and whose dependence profiles the statute judges to be limited against the schedule III baseline.

Tramadol supplies the most recent full illustration of a substance entering the schedules from outside through the administrative findings process. The synthetic opioid analgesic was approved by the Food and Drug Administration in 1995 under the trade name Ultram and remained a noncontrolled substance under federal law for nearly two decades, during which its prescribing expanded and abuse reports accumulated, including evidence that users were substituting tramadol for hydrocodone products. After evaluation, the Drug Enforcement Administration published a final rule placing tramadol in schedule IV, effective August 18, 2014. The placement recognized a currently accepted medical use for moderate to moderately severe pain, a low potential for abuse relative to the schedule III population, and limited dependence liability relative to that same baseline. The tramadol story also demonstrates the feedback loop between schedules: the same year the agency raised hydrocodone combination products to schedule II, it brought tramadol into schedule IV, adjusting two different points on the same gradient in response to related abuse patterns.

Schedule V is the floor of the system and the schedule whose findings come closest to describing ordinary pharmacy practice. The substance must have a low potential for abuse relative to the substances in schedule IV, a currently accepted medical use in treatment in the United States, and abuse that may lead to limited physical dependence or psychological dependence relative to the substances in schedule IV. The population consists largely of preparations containing limited quantities of specified narcotics: cough preparations containing not more than two hundred milligrams of codeine per one hundred milliliters, exemplified by Robitussin AC, which trace back to the original statutory listings, and the antidiarrheal preparations Lomotil and Motofen, which contain small quantities of diphenoxylate and difenoxin respectively. Pregabalin, sold as Lyrica, joined schedule V through a Drug Enforcement Administration final rule issued on July 28, 2005, shortly after the drug’s approval for neuropathic pain. The agency cited clinical trial evidence that pregabalin produced acute euphoric and positive psychic effects resembling those of the benzodiazepines diazepam and alprazolam, which sit one schedule higher, and placed the new medicine at the lowest control tier rather than leaving it unscheduled, because the schedule V findings require only a low relative potential for abuse rather than none.

Three Doors In: How a Scheduling Proceeding Begins

Knowing what the schedules require answers only half the operative question. The other half is procedural: who may ask that a substance be added, moved, or removed, and what happens once the asking begins. The Controlled Substances Act answers with a divided and deliberate allocation. The statute names three doors into the process and one gatekeeper who stands at the second door, and it assigns the Attorney General and the Secretary of Health and Human Services different halves of the final decision. Congress delegated the Attorney General’s scheduling power to the Drug Enforcement Administration, so in practice every step attributed to the Attorney General is carried out by the DEA Administrator, but the allocation of authority the statute draws is the same whether the name on the order is the Attorney General or the Administrator. Understanding that allocation is the key to understanding why scheduling proceedings unfold the way they do, why they take so long, and why the health agency’s voice can stop a scheduling action cold while the law enforcement agency’s voice cannot force one through.

The three ways a scheduling proceeding begins are stated in a single sentence of 21 U.S.C. 811(a). The Attorney General may begin a proceeding to add a substance to a schedule, transfer it between schedules, or remove it from the schedules on his own motion, at the request of the Secretary, or on the petition of any interested party. The own-motion route covers the DEA’s own surveillance and investigation work. The agency monitors medical, scientific, and law enforcement information about drugs that are not yet controlled, and when its own evidence suggests a control question, the Administrator may start the machinery without waiting for anyone else to ask. The Secretary’s request covers the reverse direction of government initiative. When the health agency concludes that a substance poses a medical or scientific problem that scheduling might address, the Secretary can ask the Attorney General to take up the question. The third route is the most open. Any interested party, which includes patients, physicians, manufacturers, trade groups, advocacy organizations, researchers, and ordinary citizens with a demonstrable interest, may file a petition asking that a proceeding be initiated to control, transfer, or remove a substance.

What must a scheduling petition contain before the DEA accepts it for filing?

Any interested party may petition: patients, physicians, manufacturers, trade groups, advocacy organizations, researchers, and ordinary citizens with a demonstrable interest. The petition must state the proposed rule, the petitioner’s interest, and the supporting facts and scientific evidence. Acceptance for filing obliges the DEA to consider the request, not to grant it.

Petitions do not enter a waiting room where they sit until someone remembers them. The regulation at 21 CFR 1308.43 prescribes the form of a petition and the track it follows once it arrives. A petition must set out the proposed rule, state the petitioner’s interest in the matter, and include a reasonably concise statement of the facts the petitioner relies on, including a summary of any relevant medical or scientific evidence known to the petitioner. Within a reasonable period after receiving the petition, the Administrator must notify the petitioner whether the petition has been accepted for filing, and if it is not accepted, must explain why. A petition that fails the formal requirements can be amended and resubmitted. Once a petition is accepted for filing, the Administrator may deny it within a reasonable time if the grounds the petitioner offers are not sufficient to justify initiating proceedings. The petition process therefore gives outside voices a formal right of entry, but it does not give them a right of command. The DEA evaluates the petition against its own evidentiary standards, and a petition that cannot support a finding of abuse potential or that fails to engage the eight statutory factors will end with a denial letter, not a rulemaking.

Petitions arrive from both directions of the scheduling question, and that symmetry reveals something about the process. Manufacturers and researchers petition to move substances to less restrictive schedules or off the schedules entirely, because scheduling carries registration, security, quota, and recordkeeping burdens that shape whether a product can be developed, prescribed, or studied. Advocacy groups and medical organizations petition to reschedule substances whose medical value they believe the schedules misstate. Law enforcement voices and community organizations petition to control new substances whose abuse they are witnessing. A denied petition is not the end of the road either. The petitioner may submit additional evidence, amend and resubmit, or, in the case of a petition denial that constitutes a final decision, seek review in court. The petition route therefore functions as the public’s permanent invitation to reopen a settled question, which matters because the statute contemplates that scientific knowledge and patterns of abuse change over time, and a scheduling decision that was correct when made can become wrong as the evidence moves.

The Health Agency Veto and the Formal Hearing

Once a proceeding begins by any of the three routes, the evaluation sequence takes center stage, and it is where Congress made its most unusual choice. Before initiating proceedings under subsection (a) to control a substance or to remove one entirely from the schedules, and after gathering the necessary data, the Attorney General must request from the Secretary a scientific and medical evaluation and a recommendation as to whether the substance should be controlled or removed. The evaluation is not an informal consultation. The statute directs the Secretary to consider the factors listed in paragraphs (2), (3), (6), (7), and (8) of subsection (c), together with any scientific or medical considerations involved in paragraphs (1), (4), and (5) of that subsection. Those are the factors that concern pharmacological effect, the state of scientific knowledge, public health risk, dependence liability, and precursor status, which is to say the factors that turn on science and medicine rather than on law enforcement observation. The Secretary’s recommendations must include a view on the appropriate schedule, if any, under which the substance should be listed. The evaluation and recommendations must be made in writing and submitted to the Attorney General within a reasonable time.

Why does the health agency’s binding power run in only one direction?

The Secretary’s recommendations bind the Attorney General as to scientific and medical matters, and if the Secretary recommends that a substance not be controlled, the Attorney General shall not control it. The binding force runs in that one direction only. The health agency can stop control but cannot compel a placement, a rescheduling, or a removal.

The sentence that gives the health agency its veto comes next. The recommendations of the Secretary are binding on the Attorney General as to scientific and medical matters, and if the Secretary recommends that a substance not be controlled, the Attorney General shall not control the substance. The direction of that binding effect matters enormously, and it is one-directional. If HHS concludes that the science does not support control, the Attorney General has no lawful path to schedule the substance, no matter what law enforcement considerations favor control. The health agency holds a veto, but not a pen. The statute gives the Secretary no power to compel the Attorney General to schedule a substance the health agency favors controlling, and it gives the Secretary no authority over the law enforcement factors, abuse patterns, and treaty obligations that the Attorney General weighs independently. The binding force attaches to scientific and medical findings only, and the veto attaches only to a recommendation against control. What Congress built, in effect, is an asymmetry: medical science can stop the machinery of control, but it cannot start it, and law enforcement can start the machinery but cannot finish the job without the health agency’s scientific concurrence.

In practice the Secretary’s evaluation is prepared through the Food and Drug Administration, which conducts the eight-factor analysis with data and analysis supplied by the National Institute on Drug Abuse. The HHS recommendation reaches the DEA as a written document that walks through each factor and states a scheduling conclusion. The DEA then applies its own independent judgment to the factors within its own competence, principally the law enforcement factors such as actual or relative potential for abuse, history and current pattern of abuse, and the scope, duration, and significance of abuse. The division of labor between the two agencies inside the evaluation itself is worth stating plainly, because the statute draws it with unusual care. The Secretary examines the five science-heavy factors in full and addresses the scientific and medical dimensions of the remaining three, while the Attorney General, through the DEA, develops the law enforcement evidence on abuse potential, abuse patterns, and the scope and significance of abuse. The two analyses then meet at the substantial-evidence threshold. A 2024 memorandum from the Department of Justice’s Office of Legal Counsel refined the timing of the binding effect: the HHS scientific and medical findings bind the DEA until the DEA issues its notice of proposed rulemaking, and the DEA must give those findings significant deference throughout the rulemaking itself, while new data presented through comments or hearing testimony may also enter the record and be weighed alongside the HHS findings. The veto is absolute when HHS recommends against control, but the rest of the process is a balancing of two agencies’ evidence rather than a rubber stamp on either one. If the Attorney General determines that the HHS evaluation and all other relevant data constitute substantial evidence of potential for abuse such as to warrant control, or substantial evidence that a substance should be removed entirely from the schedules, the Attorney General initiates proceedings for control or removal. That substantial-evidence standard is the threshold between evaluation and rulemaking. It does not require certainty or consensus; it requires enough evidence to warrant putting the question to the formal process.

The formal process is formal in a technical sense that shapes everything about how scheduling hearings look. Section 811(a) provides that the Attorney General’s scheduling rules shall be made on the record after opportunity for a hearing, pursuant to the rulemaking procedures prescribed by subchapter II of chapter 5 of title 5. That cross-reference points to sections 556 and 557 of the Administrative Procedure Act, the formal adjudication-style hearing provisions. Scheduling rulemaking is therefore not the ordinary notice-and-comment process in which an agency reads comments and issues a rule. It is an on-the-record proceeding in which interested persons may present evidence and argument before an administrative law judge, with the opportunity for cross-examination, and in which the final decision must rest on the evidence in the hearing record. This is why scheduling actions can stretch across years and produce thousands of pages of testimony. The statute chose a hearing-centered process because control of a substance restricts the lawful manufacture, distribution, and research of a drug, and Congress wanted that restriction to be supported by an evidentiary record that could be tested adversarially.

Once the DEA decides to initiate proceedings, 21 CFR 1308.43 lays out the choreography. The Administrator publishes a general notice in the Federal Register announcing the proposed rulemaking. That notice must state the legal authority under which the rule is proposed, set out the proposed rule itself, and include a statement of the time, place, and nature of any hearing that will be held if a hearing is requested. The notice also invites interested persons to file written comments or objections during a designated period, and any interested person may request a hearing or, alternatively, file a waiver of the opportunity for a hearing together with a written statement of that person’s position on the matters of fact and law involved. The regulation requires that at least thirty days pass between publication of the notice and the commencement of any hearing. The hearing, when one is held, is conducted before an administrative law judge, who hears testimony, receives evidence subject to rules of relevance and materiality, and ultimately certifies the record and issues a recommended decision. Before that recommendation becomes final, interested participants may file exceptions to it, setting out their objections in writing, and the Administrator considers those exceptions along with the certified record. The Administrator then issues the final order, which may adopt, modify, or reject the judge’s recommendation. The order is published in the Federal Register and becomes effective thirty days after publication unless the Administrator finds that public health or safety conditions require an earlier effective date. Because the rule must be made on the record, the final order must find support in the evidence the hearing produced, and it is that record that a reviewing court examines if the order is challenged. If every interested person waives the hearing or is deemed to have waived it by failing to appear, the Administrator may cancel the scheduled hearing and proceed directly to a final order on the written record, which is the statute’s answer to the problem of an uncontested proceeding.

Two features of this hearing opportunity deserve emphasis because they are frequently misunderstood. First, the hearing is available to anyone whose interests may be affected, not only to the government and not only to the pharmaceutical industry. Any interested person may file a notice of intention to participate, and the statute’s whole design assumes that researchers, physicians, patient advocates, and manufacturers will appear with evidence and witnesses. The hearing is adversarial in form precisely so that the factual premises of scheduling, especially the scientific and medical premises on which the HHS evaluation rests, can be probed rather than accepted on paper. Second, the hearing opportunity is not optional window dressing. Because the rule must be made on the record, a final order issued after a hearing request must be grounded in what the hearing record contains. Interested persons who waive the hearing and submit only written statements have those statements made part of the record, but the regulation gives them less weight in light of the absence of cross-examination. The system thus rewards participation in the hearing itself and treats the hearing as the main engine of fact-finding rather than as a courtesy.

The procedure does not end at the agency door. Section 877 of title 21 provides that any person aggrieved by a final scheduling decision may obtain judicial review in the United States Court of Appeals for the District of Columbia or in the circuit where the person’s principal place of business is located, by filing a petition within thirty days after notice of the decision. Findings of fact by the Attorney General, if supported by substantial evidence, are conclusive on review. That deferential standard of review is the reason the hearing record matters so much. Because the court will not reweigh the evidence, the parties who want to challenge a scheduling outcome must build their case inside the agency proceeding itself, in testimony and exhibits that become part of the record, rather than hoping to persuade a court from scratch later. The petitioners, the commenters, and the hearing participants are therefore not merely spectators at an agency ritual. They are constructing the only record that will ever be reviewed.

Read as a whole, then, the procedure answers the section’s title in a sentence. The Attorney General initiates, acting through the DEA, on the agency’s own motion, at the Secretary’s request, or on an outsider’s petition that survives the Administrator’s review. The Secretary evaluates, and the Secretary’s negative recommendation ends the matter. The rulemaking that follows, when it follows, is conducted on the record with a hearing before an administrative law judge open to any affected interest, and the final order that emerges must stand on the evidence that hearing produced. No single office decides. The enforcement agency proposes and disposes, but the health agency can veto, and the public can force a hearing in which every premise is tested. That division was not an accident of drafting. Congress split the pen and the veto between the two agencies whose expertise the question required, and it put the proceeding itself on the record so that the division of authority would be visible in the evidence, not merely in the statute’s text.

The Eight Factors That Decide Every Placement

The procedure described so far answers who decides and how. The eight factors answer what the decision must be about. No substance lands on a federal drug schedule because someone in a government office decided it looked dangerous. Every addition to the schedules, every transfer from one schedule to another, and every removal from control must run through a fixed checklist written into section 201(c) of the Controlled Substances Act, codified at 21 U.S.C. 811(c). The provision commands the Attorney General, who acts through the Drug Enforcement Administration, to consider eight named factors with respect to each drug or other substance proposed for control or removal. The list applies to the full life cycle of a controlled substance. A petition to schedule a new drug, a move to tighten or loosen the controls on one already listed, and a plea to take a substance off the schedules entirely all pass through the same eight inquiries.

Why don’t the eight factors settle a scheduling decision on their own?

The Attorney General must consider: actual or relative abuse potential; scientific evidence of pharmacological effect; the state of current scientific knowledge; the history and current pattern of abuse; the scope, duration, and significance of abuse; public health risk; psychic or physiological dependence liability; and whether the substance is an immediate precursor of an already controlled substance.

The eight factors do not score themselves. The statute assigns the scientific and medical work to the Department of Health and Human Services, where the Food and Drug Administration prepares a detailed eight-factor analysis and sends a scheduling recommendation back to the enforcement agency. Under section 811(b), the Secretary’s recommendations bind the Attorney General on scientific and medical questions. If Health and Human Services recommends against control, the DEA cannot schedule the substance. When the recommendation favors control, or favors a particular schedule, the DEA remains responsible for the final call: it must decide whether the facts and all other relevant data add up to substantial evidence of abuse potential sufficient to warrant control, and it must make the schedule findings spelled out in section 812(b). The two agencies therefore examine the same eight factors from different sides of the building. Health and Human Services weighs factors two, three, six, seven, and eight in full, plus whatever scientific or medical considerations sit inside factors one, four, and five; the DEA reviews that evaluation, weighs the law-enforcement and abuse data, and conducts its own eight-factor analysis across the entire record. Public notice and comment precede the final rule, and the resulting decision can face judicial review.

The analyses themselves have become part of the public record. Both the DEA and Health and Human Services eight-factor documents are posted to the public docket for each scheduling action, with the Assistant Secretary’s recommendation letter and supporting data available for inspection. When the substance under review is a new medicine moving through approval at the Food and Drug Administration, the abuse-related data the drug sponsor submitted during safety and efficacy review feed directly into the analysis, so the eight factors draw on the same evidence base as the approval decision. The transparency of the docket is what allows manufacturers, medical societies, patient groups, and reviewing courts to test each factor’s treatment rather than argue about the outcome in the abstract.

The first factor asks about the substance’s actual or relative potential for abuse, and the word relative carries most of the weight. Abuse potential is not measured against an abstract scale. It is measured against drugs already on the schedules. When the Food and Drug Administration evaluates a new compound, it compares the compound’s abuse-related properties with those of one or more pharmacologically similar substances that are already scheduled and have recognized abuse potential, often substances that are structurally similar or indicated for the same therapeutic purpose. Animal studies test whether laboratory animals find the drug reinforcing, how much effort they will expend to obtain more of it, and whether they will self-administer it when given the chance. Human abuse potential studies put the question to people with experience using drugs in the same class and record how much they report liking the new compound compared with a known drug they know. The result is a comparative judgment: this stimulant looks more like cocaine or more like a weak over the counter product, this opioid behaves like oxycodone or like something far milder.

Congress did not define abuse in the statute, but the legislative history gives the DEA a working set of signs to look for. The enforcement agency considers whether people are taking the drug in amounts large enough to create a hazard to their own health or to the safety of others and the community, and whether the drug is being diverted in significant quantities from legitimate channels into illicit ones. Those two inquiries split the first factor’s work in a useful way. The first captures harm, the second captures leakage. A medicine that heals in prescribed doses can still carry a high abuse potential if pills flood the street market, and the factor reaches that case. For substances with no medical use at all, the first factor tends to be answered by street-level evidence: seizure quantities, overdose reports, and the pharmacological company the substance keeps.

The second factor looks for scientific evidence of the drug’s pharmacological effect, with the candid qualification “if known.” Congress wrote that qualifier because the scheduling system must sometimes act before the science is complete. New synthetic drugs appear on the illicit market faster than laboratories can map their full pharmacology, and a scheduling decision cannot wait for a finished picture of every receptor interaction. When the evidence exists, it covers chemistry, structure, receptor targets, and pharmacokinetics: how the compound acts on the central nervous system, whether it mimics a stimulant, a depressant, or a hallucinogen, and how quickly it takes effect and wears off. Pharmacological similarity does particular work here. A novel compound whose structure and receptor profile closely track those of a scheduled drug can be assessed as carrying similar abuse potential even when direct evidence about the new compound is thin. That is how the machinery handles novelty. It reasons from resemblance, and it says so openly in the published analyses, which describe the degree of structural and pharmacological kinship and what that kinship implies for how humans will experience the drug. The synthetic cannabinoid FUB-AMB, a novel psychoactive substance that appeared on the illicit market with almost no human use history, illustrates the pattern. The DEA’s published eight-factor analysis of that compound leaned on factors one and two: pharmacological resemblance to already controlled synthetic cannabinoids supplied the abuse-potential comparison, and the available pharmacology stood in for a human record that did not yet exist. When factors four and five are thin because the drug is new, the statute still functions, but the decision’s weight shifts toward the laboratory and the comparison set.

The third factor sweeps wider. It calls for the state of current scientific knowledge regarding the drug or other substance, which is the statute’s way of saying that no scheduling decision should rest on a cherry-picked study. The published eight-factor documents gather the whole research record as it stands: published literature, preclinical and clinical data, and, where a drug has been developed as a pharmaceutical, the abuse-related data the sponsor submitted to the Food and Drug Administration during safety and efficacy review. The factor matters most at the margins. When pharmacology is well mapped and abuse data are rich, it confirms the case. When knowledge is fragmentary, the factor forces the government to admit the gaps, and courts reviewing scheduling decisions have treated the completeness and candor of that record as part of what makes the decision lawful. A thin research record does not block scheduling; it changes the character of the decision, and factor three is where that change gets documented.

Factors four and five form a pair, and the statute’s decision to separate them signals that history and magnitude are different questions. The fourth factor covers the history and current pattern of abuse: how long people have been misusing the substance, in what settings, through which routes of administration, and whether use is rising, falling, or holding steady. The fifth covers the scope, duration, and significance of abuse: how many people are involved, how long the problem has persisted, and how serious its consequences have been. Together they demand epidemiological evidence, and the agencies have learned where to find it. They consult emergency department data, poison control reports, addiction treatment admissions, overdose mortality figures, seizure and trafficking intelligence, and national surveys of drug use. The quality of that evidence decides close cases.

The clearest modern illustration of factors four and five in action is the DEA’s decision to move hydrocodone combination products from Schedule III to Schedule II, announced in a final rule published in August 2014 and effective that October. Hydrocodone had an accepted medical use and had sat in the less restrictive schedule because the products combined it with other ingredients such as acetaminophen, which the original scheduling scheme treated as a partial check on abuse. By the early 2010s that check had failed. The agency found that the products were being widely diverted and abused at rates comparable to oxycodone products, which already sat in Schedule II. About 137 million prescriptions for hydrocodone combinations were dispensed in 2013, and the agency tied the products to large numbers of overdose deaths and a heavy flow of patients into addiction treatment. The eight-factor record did the decisive work. Actual abuse was widespread and long-standing, the pattern matched a Schedule II opioid, and the scope was measured in millions of prescriptions and thousands of deaths. The agency’s formal conclusion tracked the statutory formula: the facts and all other relevant data constituted substantial evidence of high potential for abuse. The schedule changed because the epidemiology moved, not because the molecule changed.

The same machinery runs in the other direction. In 2023, the Department of Health and Human Services completed a full eight-factor scientific and medical evaluation of marijuana in response to a presidential request to review its scheduling, and the Food and Drug Administration recommended moving it from Schedule I to Schedule III. The evaluation walked the identical statutory path: relative abuse potential measured against drugs in the other schedules, the pharmacological record, the history and scale of use, the public health picture, and dependence liability. The episode demonstrates that the eight factors are not a ratchet that only tightens. They are the required analytic frame for loosening control as well, and a rescheduling recommendation must earn its way through each of the eight inquiries just as a scheduling one does. The longer procedural record of the cannabis proceedings, from the 1972 petition through the 2016 denials and the 2024 proposed rulemaking, is taken up later in this article.

The sixth factor asks what risk, if any, the substance poses to public health, and the cautious phrasing matters. Congress did not assume every controlled substance endangers the public. The factor asks the question and accepts that the answer might be little or none. In practice the inquiry covers harms that spill past the individual user: emergency department surges, infectious disease transmission tied to injection, impaired driving, violence in trafficking markets, and the burden on treatment systems. Public health risk has tipped close decisions in both directions. For a medicine with genuine therapeutic value, the factor weighs the community damage of diversion against the benefit of availability. For an illicit substance with no medical use, it measures the aggregate toll. The hydrocodone record shows the factor pulling toward tighter control when a widely prescribed drug generates a public health burden comparable to the most tightly controlled opioids. The same factor can also supply the factual backbone of a decision to keep a substance off the schedules, since a thin or speculative showing of community harm undermines the case for control.

The seventh factor addresses psychic or physiological dependence liability, and the two adjectives describe different hazards. Psychic dependence, in the older terminology of the statute, means the craving, the compulsive seeking, and the continued use despite harm; physiological dependence means tolerance and the withdrawal syndrome that follows when use stops. Scheduling practice treats both as serious but distinct, and the distinction shows up in the schedule definitions. Schedule II status turns in part on use potentially leading to severe psychological or physical dependence, while the lower schedules describe progressively lighter dependence profiles. The dependence factor has carried real weight in pharmaceutical scheduling. When the DEA evaluates a newly approved medicine, animal and human studies of withdrawal and of continued drug-seeking after access is restricted supply much of the evidence, supplemented by postmarketing reports once the drug reaches the market. A drug that produces a punishing withdrawal syndrome or that patients keep seeking long after the medical need has passed will land higher on the schedules than a pharmacological cousin with a gentler profile, even when the two share a therapeutic purpose.

The eighth factor looks upstream. It asks whether the substance is an immediate precursor of a substance already controlled, which gives the government a tool against the supply chain rather than the finished product. A chemical that sits one step away from a controlled drug, and whose principal destination is conversion into that drug in clandestine laboratories, can be scheduled in its own right to choke off manufacture at the source. The word immediate does real interpretive work. It excludes distant raw materials with broad legitimate industrial uses and reaches the compounds that exist primarily to become the controlled substance. This is the factor that lets scheduling follow the chemistry of illicit production rather than waiting for each new finished drug to appear on the street.

Removal from the schedules runs on the same eight factors, in reverse. The statute directs the Attorney General to consider the factors with respect to each substance proposed for removal, and a removal requires a finding that the substance no longer meets the requirements for inclusion in any schedule. A drug whose abuse has faded, whose science has clarified a benign profile, or whose dependence liability was overstated on a thin record can be argued out of control on the very grounds that put it there. The burden is real, since the government must make an affirmative finding across the same eight inquiries, but the mechanism exists precisely because scheduling decisions are supposed to track the evidence rather than freeze the judgments of an earlier decade.

The class-wide control of fentanyl-related substances shows how the emergency lane interacts with the eight-factor machinery. When fentanyl analogues began appearing in the illicit supply faster than individual scheduling actions could respond, the DEA used its temporary authority in February 2018 to place the structural class of fentanyl-related substances in Schedule I, the first administrative class-wide scheduling action. Temporary orders are capped in duration and do not require the full eight-factor proceeding; they exist for imminent hazards that cannot wait for the ordinary process. The eight factors govern the administrative route, with its scientific evaluation, its binding HHS findings, its public comment, and its judicial review. The emergency route exists for cases where speed overrides that deliberate process, and the temporary order’s two-year term defines its boundary: it holds the line while the permanent proceeding runs its course, and it expires on its own clock.

Taken together, the eight factors ask one sustained question in eight voices: what is this substance, how do people misuse it, how many people are harmed, and how does it compare with the drugs the schedules already contain. The first factor sets the comparison, the second and third supply the science, the fourth and fifth supply the epidemiology, the sixth measures the community toll, the seventh tests the grip the drug takes on its users, and the eighth looks one step up the chemical ladder. No single factor decides. The DEA’s published decisions weigh the totality, the Health and Human Services evaluation binds the science, and the public record shows its work. That structure is why scheduling disputes so often turn into fights about the factors themselves: whether the pharmacological resemblance was overstated, whether the abuse statistics measured the right population, whether the dependence evidence came from the right species of study. The eight factors do not remove judgment from drug scheduling. They organize it, force it into the open, and give everyone watching, from manufacturers to patient groups to reviewing courts, a common vocabulary for arguing about where a substance belongs.

Emergency Scheduling and Treaty Obligations at the Outer Edges

The eight factors describe the ordinary machinery, and the ordinary machinery is deliberately slow. A substance moves through petitions, a scientific and medical evaluation prepared by the Secretary of Health and Human Services, an opportunity for hearings, a formal order of the Attorney General, and the possibility of judicial review. The design assumes that the right classification of a drug is a question for deliberation, not for speed. In 1984, Congress decided that deliberation had a cost. Chemists working for clandestine laboratories were introducing new substances faster than the scheduling process could absorb them, and the government found itself without legal authority to seize, prosecute, or restrict drugs that were already killing people. The answer, enacted through the Comprehensive Crime Control Act of 1984 as a new subsection of 21 U.S.C. section 811, was an emergency lane: temporary scheduling to avoid an imminent hazard to the public safety. It has never been a shortcut to permanent control. It is a holding action, and every one of its design features reflects the fact that Congress meant it to remain temporary.

How long can temporary scheduling last?

A temporary order expires at the end of two years from issuance. If scheduling proceedings are pending, the Attorney General may extend the placement for up to one year, and the order is vacated when those proceedings conclude. The 1984 limits were shorter, one year plus a six-month extension, lengthened by the Synthetic Drug Abuse Prevention Act of 2012.

The provision lives at section 811(h). Its trigger is a finding by the Attorney General that placing a substance in Schedule I on a temporary basis is necessary to avoid an imminent hazard to the public safety. That finding unlocks a power that exists nowhere else in the scheduling apparatus. The Attorney General may act by order, without regard to the ordinary requirements of subsection (b) that route the scientific and medical evaluation through the Secretary of Health and Human Services. The bypass is not total. When issuing a temporary order, the Attorney General is required to consider only three of the eight statutory factors that govern ordinary scheduling: the history and current pattern of abuse of the substance, the scope, duration, and significance of abuse, and the risk to the public health. Those three factors, the provisions paragraphs (4), (5), and (6) of subsection (c), correspond to the evidence that can be gathered quickly: what is happening on the street, how widespread the abuse has become, and what harm is resulting. The Attorney General must also transmit notice of a proposed order to the Secretary of Health and Human Services and take into consideration any comments the Secretary submits in response. The health officials are consulted but do not hold a veto, a deliberate inversion of the normal relationship in which the Secretary’s scientific recommendation binds the Attorney General on medical questions.

Two eligibility conditions narrow the field further. A temporary order may issue only if the substance is not listed in any other schedule, and only if no exemption or approval for the substance is in effect under section 505 of the Federal Food, Drug, and Cosmetic Act. An approved medicine, in other words, can never be pulled into Schedule I by emergency fiat. And the process itself carries a waiting period. No temporary order may be issued before thirty days have passed from both the publication in the Federal Register of a notice of intent stating the grounds for the order and the transmittal of that notice to the Secretary of Health and Human Services. The emergency lane is therefore not instantaneous. It compresses the years of ordinary rulemaking into weeks, but it preserves a period for the public and the health authorities to be informed and to respond.

The authority formally belongs to the Attorney General, but in practice it is exercised by the Administrator of the Drug Enforcement Administration under a standing delegation of the Justice Department’s scheduling powers. Every notice of intent and final order under the provision names the DEA Administrator as the acting official, and the agency’s Office of Diversion Control assembles the evidentiary record that supports the imminent hazard finding. That record draws on poison control center reports, hospital data, law enforcement seizures, and surveys of abuse patterns, which is why the three factor limitation works as an evidentiary shortcut rather than a mere abbreviation. The factors the statute preserves, history and pattern of abuse, scope and significance, and public health risk, are exactly the factors that can be documented from field reporting within weeks. The factors it drops, scientific evidence of pharmacological effect, the state of current scientific knowledge, dependence liability, and the accepted medical use inquiry, are the ones that require laboratory studies, literature reviews, and the full HHS evaluation process. Congress made a deliberate trade. The emergency order gets to market on street level evidence, and the question of what the drug actually does to the body is deferred to the permanent proceeding.

The most important constraints are temporal. A temporary scheduling order expires at the end of two years from the date of issuance. If ordinary scheduling proceedings under subsection (a)(1) are pending with respect to the substance, the Attorney General may extend the temporary placement for up to one additional year, and the order is vacated upon the conclusion of those proceedings. Congress built a hard deadline into the power precisely because the order issues with reduced procedural protection. A temporary scheduling order is also, by statute, not subject to judicial review. That exemption from review is the clearest signal of the provision’s character. Ordinary scheduling orders may be challenged in court, where judges test whether the agency’s findings rest on substantial evidence. Temporary orders escape that scrutiny entirely, which is tolerable only because they cannot endure. The original 1984 time limits were even shorter, one year with a possible six month extension, and Congress lengthened them to the present two years plus one year in the Synthetic Drug Abuse Prevention Act of 2012. The lengthening reflected a pragmatic calculus. The one year window had been written for an era when a new drug of abuse appeared occasionally and the permanent process could be completed within the emergency period. By the time synthetic cannabinoids and cathinones were arriving in waves, with each controlled chemical replaced within months by a molecular variant designed to evade the law, a single year no longer covered the full rulemaking, so the longer runway kept the control from lapsing midstream. The extension remains available only while proceedings for permanent scheduling are actually pending, and the design still contains its own release valve. If the permanent proceeding concludes and finds that the substance does not belong in Schedule I, the temporary order is vacated and control lifts, which means the emergency finding of imminent hazard is never the final word on a drug’s status.

The record of actual use shows what the mechanism was built for: waves of designer drugs that the ordinary process could not catch. The first large scale deployment came with synthetic cannabinoids sold under names like Spice and K2. In November 2010 the Drug Enforcement Administration published a notice of intent to temporarily place five synthetic cannabinoids in Schedule I, citing their high abuse potential, the absence of any recognized therapeutic use, and evidence that uncontrolled manufacture and distribution posed an imminent hazard. The final temporary order followed on March 1, 2011, covering JWH-018, JWH-073, JWH-200, CP-47,497, and cannabicyclohexanol. In February 2012, while permanent scheduling proceedings were pending, the agency extended the temporary placement under the statute’s extension provision to August 29, 2012, or until the rulemaking concluded, whichever came first. The synthetic stimulants sold as bath salts followed a similar arc. In September 2011 the agency published a notice of intent to temporarily schedule three synthetic cathinones, mephedrone, methylone, and MDPV, based on a finding that their placement in Schedule I was necessary to avoid an imminent hazard, and the final order issued the following month. Congress later placed those synthetic cannabinoids and cathinones into Schedule I permanently by statute, which is the pattern the emergency provision anticipates: the temporary order holds the line while the permanent decision is made through the full process.

If temporary scheduling is the emergency lane for control, the treaty provision of the same statute is the ceiling that limits how far control can be undone. Section 811(d)(1) addresses the situation in which control of a substance is required by United States obligations under international treaties, conventions, or protocols in effect on October 27, 1970. When that condition is met, the Attorney General shall issue an order controlling the drug under the schedule he deems most appropriate to carry out those obligations, and he does so without regard to the findings ordinarily required and without regard to the ordinary scheduling procedures. The word shall makes the duty mandatory. The statute does not invite the Attorney General to weigh whether treaty required control is wise; it commands him to impose it, and it excuses him from the domestic deliberative apparatus in the process. For psychotropic substances governed by the later Convention on Psychotropic Substances of 1971, the statute carries a separate procedure, added by the Psychotropic Substances Act of 1978. When the Secretary of State receives word from the United Nations that the World Health Organization is considering adding, transferring, or deleting a substance under that convention, the notice passes to the Secretary of Health and Human Services, who publishes it in the Federal Register, solicits public comment, and prepares the medical and scientific evaluations that the United States transmits back through the Secretary of State. The United States participates in the international scheduling decision, and domestic control follows from the international outcome. The provision also handles the situation in which domestic law is already ahead of the treaty. If the Secretary determines that existing controls under the Controlled Substances Act and the Federal Food, Drug, and Cosmetic Act already meet the requirements of the convention schedule at issue, no new order is needed, though the Secretary may still recommend ordinary scheduling proceedings if more stringent controls seem warranted. If existing controls fall short and the Secretary concurs in the international scheduling decision, he recommends that the Attorney General initiate scheduling under the appropriate domestic schedule. The effect is to make American drug scheduling partly downstream of decisions made in Geneva. The scientific evaluation happens first at the World Health Organization, the member states vote, and the domestic machinery then implements the result.

The consequence is that descheduling is not a purely domestic decision. The Controlled Substances Act allows the Attorney General to remove a substance from the schedules or transfer it between them, and the factors in subsection (c) guide that judgment. But where an international convention obligates the United States to maintain control of a substance, the domestic analysis runs into a floor set by treaty. The Attorney General may not simply strip away control and leave the United States in breach of its obligations. The Single Convention on Narcotic Drugs of 1961, which was in force when the Controlled Substances Act was enacted, is the principal treaty at issue. It sorts narcotic drugs into international schedules with specified control measures, and the United States implemented those obligations through its domestic schedules. A domestic decision to remove a treaty covered substance entirely would collide with the treaty command to keep it under control.

The history of marijuana scheduling shows the constraint in action and also shows that its scope has been contested across decades. In 1972, when a reform organization petitioned the Bureau of Narcotics and Dangerous Drugs, the predecessor of the Drug Enforcement Administration, to deschedule cannabis or move it to Schedule V, the Bureau’s director refused even to accept the petition, taking the position that the Single Convention barred the action. Litigation followed, and in 1977 the Court of Appeals for the District of Columbia Circuit, in NORML v. DEA, 559 F.2d 735, reasoned that treaty obligations prevented marijuana from being placed in a schedule less restrictive than Schedule II. The court’s reading made the treaty a structural limit on the agency’s domestic discretion. In 1988, when the agency’s own administrative law judge recommended transferring marijuana to Schedule II, the Administrator rejected the recommendation and kept the substance in Schedule I. When the agency denied a new round of rescheduling petitions in August 2016, the denial letter stated that in view of United States obligations under international drug control treaties, marijuana could not be placed in a schedule less restrictive than Schedule II, and the accompanying analysis explained the treaty reasoning in detail. Across four decades, the agency repeatedly treated the Single Convention as a barrier to any domestic move below Schedule II.

That reading has not gone unchallenged, and the dispute illustrates the second half of the treaty constraint: it limits how far control can be relaxed, but it does not dictate every domestic classification above the minimum. The statute directs the Attorney General to choose the schedule he deems most appropriate to carry out the obligations, which preserves executive judgment about what the treaty actually requires. A later legal analysis by the Justice Department’s Office of Legal Counsel concluded that neither the Single Convention nor the Controlled Substances Act required keeping marijuana in Schedule I or Schedule II, a reading that senators, state department officials, and cannabis policy advocates each advanced in public debate, while other lawmakers insisted that anything below Schedule II would violate the convention. The disagreement turns on the details of the convention’s control regimes, on how the United Nations removed cannabis from one of the convention’s schedules in 2020, and on what the domestic schedules demand in practice. The mechanism of the constraint is settled: a treaty obligation of control overrides the domestic factors and procedures, and the Attorney General must act. The exact boundary of what the obligation requires in any given case is the point on which administrations, courts, and legislators have differed.

The treaty obligation also operates as a floor rather than a ceiling, which is why it matters most in descheduling debates. The statute asks for the schedule the Attorney General deems most appropriate to carry out the obligations, so nothing stops the United States from maintaining controls stricter than the convention’s minimum. The treaty commands at least this much control; domestic law may add more. Removing control below the required level is the move the provision forbids. That asymmetry explains the one way character of the constraint. Adding control is always permissible and can even be required. Shedding control downward is permitted only to the point where the treaty floor holds. A government that wishes to take a substance off the schedules entirely must therefore answer two questions, not one: whether the domestic factors support removal, and whether the international commitments permit it. The second question has its own procedure, its own decision makers, and its own forum, which is exactly what makes descheduling not a purely domestic decision.

The two halves of this section belong together because they answer the same underlying question from opposite directions. Temporary scheduling asks what the government may do when danger is immediate and the ordinary process is too slow, and the statute answers that it may act fast but only briefly, only in Schedule I, and only on a record of abuse and public health risk. The treaty provision asks what the government may not undo when the United States has promised the world to maintain control, and the statute answers that domestic discretion yields to international obligation, without the usual findings and without the usual procedure. One provision accelerates control, the other anchors it. Between them, they define the outer edges of the scheduling power: an emergency authority that cannot become permanent by drift, and an international commitment that cannot be erased by a domestic stroke of the pen.

The Schedule I Circularity Question and the Movement Record

The criteria, the procedure, the factors, and the outer edges are now in place. What remains is the question the system is most often asked: whether the most restrictive schedule is a trap from which no substance can argue its way out, and whether anything has ever actually moved. The Controlled Substances Act places a substance in Schedule I only when three findings are made together, and the statute states them plainly in 21 U.S.C. 812(b)(1): the substance has a high potential for abuse, it has no currently accepted medical use in treatment in the United States, and there is a lack of accepted safety for its use under medical supervision. The first finding is about danger, the third is about risk under a doctor’s care, and the middle one, the medical use finding, is where the placement of cannabis has been fought for nearly five decades. The fight has turned not on what the words mean in the abstract but on who decides whether medical use exists and what kind of proof counts.

Why do rescheduling petitioners call Schedule I circular?

NORML, MAPS, and allied researchers argue that Schedule I placement withholds the registrations, the research priority, and the supply that adequate and well-controlled studies require, and then the absence of those studies justifies keeping the placement. The loop, on this view, is that the schedule prevents the evidence that would justify leaving it.

The Drug Enforcement Administration’s answer to that question has been an evidentiary standard. In a March 1992 final rule published at 57 Fed. Reg. 10499, the agency set out a five part test for deciding whether a drug has a currently accepted medical use. The test requires that the drug’s chemistry be known and reproducible, that adequate safety studies exist, that adequate and well-controlled studies prove efficacy, that qualified experts accept the drug, and that the scientific evidence be widely available. The Court of Appeals for the District of Columbia Circuit upheld the standard in Alliance for Cannabis Therapeutics v. DEA, 15 F.3d 1131 (1994), reasoning that the Administrator had reasonably insisted on rigorous scientific proof over anecdotal evidence. At the center of the test sits the third element, the demand for adequate and well-controlled studies, which is the evidentiary standard the Food and Drug Administration applies before approving a new drug. The agency did not treat physician testimonials, isolated case studies, or clinical impressions as sufficient, a position the 1992 rule stated explicitly and the court sustained.

That demand for clinical trial evidence is where NORML, MAPS, and allied researchers locate a circularity. Conducting controlled clinical trials on a Schedule I substance requires a researcher registration from the DEA and protocol review by the Food and Drug Administration, and for marijuana the supply question added a further layer. The National Institute on Drug Abuse was for decades the sole authorized source of research grade marijuana, grown under contract at the University of Mississippi, and from 1999 through June 2015 non-federally funded research protocols involving marijuana also had to pass a Public Health Service review that HHS later eliminated. Organizations that pursued medical cannabis research described this arrangement as a closed loop. The Multidisciplinary Association for Psychedelic Studies, which sponsored a study of marijuana for post traumatic stress disorder, saw the National Institute on Drug Abuse refuse to supply cannabis for an FDA approved protocol in 2011; the group obtained the Public Health Service panel’s approval in 2014 and was then told the needed cannabis was not available, and the study did not receive its final approval until April 2016. The argument, as stated by NORML, MAPS, and allied researchers, is that the Schedule I placement withholds the registrations, the funding priority, and the supply that the adequate and well-controlled studies require, and then the absence of those studies is cited to justify keeping the placement. The criticism is not new. In Alliance for Cannabis Therapeutics v. DEA, 930 F.2d 936 (1991), the same court that later upheld the five part test found three of the agency’s earlier medical use factors to be arbitrary precisely because they were, in the court’s words, logically impossible for Schedule I drugs to satisfy, since widespread therapeutic use cannot exist for a substance the law forbids doctors to use.

The mechanics of that pathway matter because the circularity debate turns on how heavy each gate is. A researcher who wants to study a Schedule I substance must hold a DEA researcher registration, submit research protocols for agency review, and satisfy the physical security requirements the regulations impose on Schedule I work. The Food and Drug Administration reviews the protocol under its drug development process, and for marijuana the supply gate stood apart from every other Schedule I drug: the only lawful source was the National Institute on Drug Abuse, which contracted a single grower at the University of Mississippi, while every other Schedule I research substance could be obtained from multiple licensed manufacturers. From 1999 until June 2015, a non-federally funded marijuana protocol faced an additional Public Health Service review that no other Schedule I substance required; HHS eliminated that layer in 2015, and in December of that year the DEA announced a waiver making it easier for researchers conducting clinical trials with cannabidiol to modify protocols and obtain additional supply. The critics’ point, stated by MAPS and by NORML in comments on the rulemakings, was that each of these gates, the registration, the protocol reviews, the single source supply, and the funding priority that followed legal status, slowed or blocked the controlled trials the five part test demanded, while the test itself borrowed the pharmaceutical approval standard from a development pathway that Schedule I placement foreclosed. The 1991 ruling is the closest the courts came to naming the circularity as a legal problem. The D.C. Circuit did not dispute that the Administrator could require scientific evidence of medical use; it held that three of the agency’s eight factors were arbitrary and capricious because they demanded, as preconditions for recognizing medical use, the very conditions, general availability, pharmacopeial recognition, and widespread physician use, that Schedule I placement makes impossible. The 1992 five-part test was the agency’s answer: it dropped the availability and recognition factors, kept the demand for adequate and well-controlled studies, and the 1994 court treated the petitioners’ concession that the revision cured the earlier defect as decisive. The circularity complaint therefore survived the litigation in a narrower form. The test no longer required the impossible, but its centerpiece still required clinical trials whose conduct the placement encumbered, and the later research access reforms, the 2015 end of the Public Health Service review and the 2016 grower expansion, can be read as the agencies addressing the encumbrance without surrendering the standard.

The agencies answer that the circle is not closed, that the test is passable, and that the court agreed. In the 1994 decision the court observed that none of the five criteria was impossible for a Schedule I drug to meet and noted that the petitioners had conceded that the new standard corrected the flaws the court had identified in the earlier version. Agency defenders add that the evidentiary bar exists to protect patients from medicines whose safety and efficacy have not been established under controlled conditions, and they point to the conclusion the Food and Drug Administration reached during the 2016 rescheduling review that medical and scientific data did not yet prove marijuana safe and effective as a medicine. On that view the placement reflects the state of the evidence rather than preventing its development, and the same rulemaking process that denied rescheduling also produced the research supply reform described below.

The schedules themselves give the reply to anyone who treats a placement as permanent. The statute authorizes the Attorney General, through the DEA, to add substances and to transfer them between schedules by rule, and Congress has itself placed and reclassified substances by legislation. Movement in both directions is part of the record.

The movement downward has been most visible where a marketed medicine’s record of use under medical supervision supplied the evidence the statute asks for. Dronabinol, synthetic delta 9 tetrahydrocannabinol sold in capsule form as Marinol, began in Schedule I. The DEA issued a notice of proposed rulemaking in October 1985 to transfer the capsule formulation to Schedule II and published its final rule on July 13, 1986, at 51 Fed. Reg. 17476. Once the medicine had been marketed and its abuse potential assessed against the record of actual use, the agency moved the Marinol formulation again, to Schedule III, in 1999, leaving other formulations of the same chemical in Schedule I. Buprenorphine traveled a similar route on a different theory. The semi synthetic opioid had been placed in Schedule V, and acting on a formal recommendation from the Department of Health and Human Services, the DEA proposed moving it up two schedules to Schedule III, a final rule issued in 2002, because reports of abuse of stronger formulations abroad and the prospect of more potent tablet versions in the United States showed that the lowest schedule no longer fit even as the drug’s accepted medical use in treating opioid dependence was established.

The movement upward shows the mirror image. Hydrocodone combination products, the Vicodin and Norco formulations that pair hydrocodone with acetaminophen or other non opioid ingredients, sat in Schedule III for decades while hydrocodone itself sat in Schedule II. A physician’s 1999 petition set the machinery in motion. An FDA advisory committee voted 19 to 10 in January 2013 to recommend the transfer, HHS sent its recommendation to the DEA in December 2013, the agency published a notice of proposed rulemaking in February 2014, and the final rule issued August 22, 2014, at 79 Fed. Reg. 49661, moved the combination products to Schedule II effective October 6, 2014. The rule’s rationale was not that the medical use had vanished but that the abuse and dependence data no longer matched the Schedule III findings.

Congress has also moved substances directly by statute, sometimes creating the split scheduling that the cannabis record discusses. The Hillory J. Farias and Samantha Reid Date Rape Drug Prohibition Act of 2000, Pub. L. 106-172, directed the scheduling of gamma hydroxybutyric acid in Schedule I while the statute’s structure preserved an FDA approved sodium oxybate medicine, Xyrem, in Schedule III. The Anabolic Steroids Control Act of 1990 placed anabolic steroids in Schedule III by legislation. And substances can leave the schedules entirely. Propoxyphene, marketed as Darvon and Darvocet in Schedule IV, followed its market exit off the schedules: after the FDA asked in November 2010 that propoxyphene products be withdrawn from the United States market because of fatal heart rhythm risks, the DEA removed the drug from control, ending the federal scheduling of a medicine that no longer had a lawful market to regulate.

The examples share a procedural posture that the cannabis record will show in its most litigated form. Congress wrote the initial schedules itself in 1970, placing named substances in each schedule through Pub. L. 91-513. Since then most movement has come through section 811 rulemaking: the Attorney General, acting through the DEA, must request a scientific and medical evaluation from the Secretary of Health and Human Services before acting, must publish a proposal, must provide an opportunity for a hearing on the record, and must make the findings the destination schedule requires. Denials of rescheduling petitions are reviewable in the Court of Appeals for the District of Columbia Circuit, which is why the cannabis petitions generated the body of case law that defines the medical use standard. The downward moves, the upward moves, the congressional placements, and the removals all passed through that machinery, and each one shows the schedules responding to evidence, whether evidence of abuse that outran the old findings or evidence of medical use and manageable risk that the old findings understated.

The cannabis record runs alongside those movements without joining them, and its procedural detail is the point. The original placement came with the statute in 1970. On May 18, 1972, NORML petitioned the Bureau of Narcotics and Dangerous Drugs, the DEA’s predecessor, to remove marijuana from control or transfer it to Schedule V. The Bureau’s director refused to accept the petition for filing on September 1, 1972, citing the Single Convention on Narcotic Drugs, and the D.C. Circuit in 1974 told the agency that treaty obligations did not excuse it from deciding the petition on its merits. After a hearing and a denial, the court remanded again in 1977 in NORML v. DEA, 559 F.2d 735, ordering the agency to follow the statute’s process, including scientific and medical review by the health department. That review, completed in 1979, concluded that marijuana should remain in Schedule I, and the DEA denied the petition again; an unpublished 1980 order from the court sent the matter back once more. The case finally reached a full evidentiary hearing in 1986, docket 86-22, after which Chief Administrative Law Judge Francis Young issued his opinion on September 6, 1988, finding that marijuana did not meet the legal criteria for Schedule I and recommending transfer to Schedule II, with the observation that the plant in its natural form was among the safest therapeutically active substances known. Administrator John Lawn rejected the recommendation in a December 1989 final order, relying on the testimony of physicians and researchers who had conducted detailed studies and on the absence of the kind of scientific evidence the statute’s findings require.

The courts then narrowed the agency’s test without ordering the transfer. In the 1991 Alliance for Cannabis Therapeutics decision the D.C. Circuit struck down three of the eight factors the agency had been using, the requirements of general availability, recognition in standard reference works, and use by a substantial segment of medical practitioners, as impossible for any Schedule I substance to meet. The agency responded with the March 1992 final rule and its five part test, and the 1994 decision upheld the denial under that revised standard. The petitions continued through the same channel. A petition filed in 1995 by former NORML director Jon Gettman was denied in 2001. A Coalition to Reschedule Cannabis petition filed in October 2002 by Gettman was denied in July 2011, the agency finding that marijuana still met all three Schedule I criteria. The D.C. Circuit upheld that denial in Americans for Safe Access v. DEA on January 22, 2013. In August 2016 the DEA denied two further petitions and, in the same announcement, adopted a new policy ending the requirement that all research marijuana come through the National Institute on Drug Abuse, opening applications for additional licensed growers. The denial kept the Schedule I placement; the grower policy addressed the supply half of the circularity complaint while leaving the evidentiary standard untouched.

The two positions the record supports can be stated without choosing between them. The public health caution position, stated by the DEA in its denials and by the Food and Drug Administration in its scientific evaluations, holds that marijuana meets the Schedule I findings because its abuse potential is high, its medical use has not been established through adequate and well-controlled studies, and its safety under medical supervision has not been accepted, and that lowering the evidentiary bar for one substance would weaken the standard that protects patients across the drug supply. The 2016 denial rested on the FDA’s medical and scientific evaluation, conducted under the section 811(b) process, which concluded that the available evidence did not establish safety and efficacy; the DEA then found each of the three statutory findings still satisfied. On this view the consistency of the denials across four decades is not stubbornness but the evidence base failing, repeatedly, to clear a bar that medicines clear when the trials exist, as Marinol’s passage from Schedule I to Schedule II to Schedule III showed. The research access position, stated by MAPS, NORML, and researchers who sought registrations and supply, holds that the Schedule I restrictions themselves, the separate DEA registration, the single source supply through the National Institute on Drug Abuse, and the added Public Health Service review that stood from 1999 to 2015, throttled the very studies the standard demands, and it points to Judge Young’s 1988 finding and to medicines like Marinol, which reached Schedule III only after being studied as a pharmaceutical, as evidence that the plant’s chemistry was never given the same path. On this view the consistency of the denials is the product of the test’s design: a standard imported from pharmaceutical development, applied to a substance barred from pharmaceutical development, with the 2016 grower expansion arriving only after the evidentiary record for the denial had already been assembled. Both positions draw on the same administrative record, and the record through August 2016 shows a placement that survived every procedural challenge while the agency steadily modified the research access rules around it.

The record after that point is stated here only through dated events, with no claim about any outcome. In August 2023 the Department of Health and Human Services issued its scientific and medical evaluation recommending that marijuana be transferred from Schedule I to Schedule III, and on May 21, 2024, the Department of Justice published a notice of proposed rulemaking consistent with that recommendation. Both are stated as actions with their dates and procedural posture, and neither is presented as a final determination. Any development after those dates should be verified against the Federal Register and the courts before being treated as the placement’s status.

A Veto Without a Pen: The Honest Account of the Pace

The record so far shows a system with real rules: conjunctive findings, an evidentiary checklist, a health agency veto, adversarial hearings, treaty floors, and a documented history of movement in both directions. The remaining question is the tempo, and the answer to it is the scheduling system’s most quoted claim. The scheduling system gives the health agency a veto and not a pen. That single asymmetry governs the pace at which substances move between the five schedules, and it explains why a substance can sit in Schedule I for decades while the science around it shifts. The Secretary of Health and Human Services can stop a scheduling action in its tracks, but the statute contains no mirror-image sentence compelling the Attorney General to schedule, reschedule, or deschedule anything. The power runs one way, and once that directionality is understood, the slow tempo of downward movement stops looking like an accident and starts looking like the product Congress designed.

The mechanics begin with a request the Attorney General cannot skip. Before initiating proceedings to control a drug or other substance, or to remove one entirely from the schedules, the Attorney General must gather the necessary data and then ask the Secretary for a scientific and medical evaluation and for recommendations on whether the substance should be controlled or removed. The statute spells out how the Secretary must build that evaluation. The Secretary considers five of the eight factors listed in subsection (c), covering scientific evidence of pharmacological effect, the state of scientific knowledge, public health risk, dependence liability, and immediate precursor status, plus any scientific or medical considerations involved in the other three, which cover abuse potential, history and pattern of abuse, and the scope, duration, and significance of abuse. The evaluation and the recommendations must be written, must be delivered within a reasonable time, and must include the Secretary’s view of the appropriate schedule, if any, under which the substance should be listed. The proceeding itself can begin from three different starting points: the Attorney General may act on his own motion, at the request of the Secretary, or on the petition of any interested party. The statute deliberately leaves the door open to manufacturers, physicians, medical societies, public interest groups, and individual citizens, any of whom can ask the enforcement agency to reconsider a placement. That openness matters, because it means the process is not closed to outsiders. But the gate still swings only where the Attorney General pushes it. A petition obliges the agency to consider the request; it does not oblige the agency to grant it.

What do subsections (i) and (j) of section 811 each govern?

Citations that stop at (h) are reading the original map. Subsection (i), on temporary and permanent scheduling of newly identified anabolic steroids, arrived through the Designer Anabolic Steroid Control Act of 2014; subsection (j), the interim final rule procedure for newly approved drugs with abuse potential, arrived in November 2015. After the 2014 and 2015 additions, the section runs from (a) through (j).

Then comes the sentence that creates the veto. The recommendations of the Secretary to the Attorney General shall be binding on the Attorney General as to scientific and medical matters, and if the Secretary recommends that a drug or other substance not be controlled, the Attorney General shall not control the drug or other substance. The binding quality attaches to the science. When the health agency, after working through the statutory factors, concludes that a substance should not be controlled, the Attorney General has no lawful path around that conclusion. The proceeding ends before it begins. That is the veto, and it belongs entirely to the health agency.

What the statute does not contain is any sentence running the other direction. Nothing in the text requires the Attorney General to place a substance in the schedule the Secretary recommends, or to act at all on a recommendation favoring a lower schedule or removal. The Secretary’s scheduling recommendation is advisory as to placement; the binding force covers only the scientific and medical findings underneath it. The Office of Legal Counsel addressed exactly this boundary during the marijuana rescheduling proceedings and concluded that the health agency’s scheduling recommendations were not binding on the Drug Enforcement Administration’s ultimate placement decision. The health agency evaluates the science and medicine, and its negative verdict kills control, but the affirmative decision of where to place a substance, and whether to revisit an old placement, remains with the Attorney General, acting through the DEA administrator under delegated authority.

The statute reinforces that control at the very next step. After receiving the health agency’s evaluation, the Attorney General decides whether the facts and all other relevant data constitute substantial evidence of potential for abuse such as to warrant control, or substantial evidence that the substance should be removed entirely from the schedules. Only if the Attorney General makes that determination does he initiate proceedings. The threshold phrase belongs to the enforcement agency, and the decision to cross it is discretionary. A favorable recommendation from the health agency toward rescheduling or removal therefore arrives at a desk where it can be respectfully received and then set aside. The statute never says it must be acted on. This is the precise sense in which the health agency lacks a pen: its science can forbid a result, but its judgment cannot command one.

One qualification keeps the picture honest. The statute does contain a single provision in which a health agency recommendation triggers mandatory action in the affirmative direction, and it concerns new drugs. When a new-drug application is submitted for a drug with a stimulant, depressant, or hallucinogenic effect on the central nervous system and abuse potential appears, the information is forwarded to the Attorney General. If the Secretary then recommends control in Schedule II, III, IV, or V, the Attorney General must issue an interim final rule controlling the drug within ninety days of receiving the recommendation or the drug’s approval, whichever is later. That interim rule takes effect immediately, without a showing of good cause, though interested persons get an opportunity to comment and request a hearing before the final rule issues. The provision shows that Congress knew how to write a mandatory command when it wanted one. It wrote that command for newly approved drugs moving onto the schedules, and it did not write one for established substances seeking to move down or off. The asymmetry is deliberate, not an oversight.

The consequence is a ratchet that tightens more readily than it loosens. A recommendation to control more strictly carries institutional momentum behind it, because the enforcement apparatus initiates the proceeding and writes the final rule. A recommendation to move downward or to remove a substance entirely faces a different reality: it asks the same enforcement apparatus to surrender authority it already holds, through a formal rulemaking on the record after opportunity for a hearing, with findings made against each schedule’s statutory criteria. Nothing compels the agency to complete that journey. The veto explains why scheduling errors of over-control persist; the missing pen explains why corrections arrive slowly, if at all.

The marijuana record illustrates the tempo. On May 21, 2024, the Department of Justice published a notice of proposed rulemaking to transfer marijuana from Schedule I to Schedule III, consistent with the health agency’s view that marijuana has a currently accepted medical use, a potential for abuse less than substances in Schedules I and II, and an abuse profile of moderate or low physical dependence or high psychological dependence. A hearing notice followed in August 2024, and through the close of that year no final rule had issued and the proposed rule had not changed the substance’s placement. The episode is stated here as a dated procedural posture, not as an outcome. What it demonstrates, on the statute’s own terms, is the unevenness of the lanes: a proposed downward move travels the slow road of formal rulemaking, while other provisions of the same statute move faster in the other direction, which is the subject of the second half of this account.

The skeptical reading treats all of this as politics wearing a lab coat, and that reading deserves an answer, because the statute does more than decorate a political choice. Start with the scientific evaluation requirement, which is not ornamental. The eight factors the Attorney General must consider in making any scheduling finding are specific and substantive: the actual or relative potential for abuse, the scientific evidence of pharmacological effect, the state of current scientific knowledge, the history and current pattern of abuse, the scope, duration, and significance of abuse, the risk to the public health, the psychic or physiological dependence liability, and whether the substance is an immediate precursor of an already controlled substance. The statute says the Attorney General shall consider these factors, and the word shall is doing real work. A scheduling rule that ignored the factors would not survive judicial review, because the statute makes the findings mandatory predicates for every placement. The formal rulemaking on the record, with opportunity for a hearing, compounds the constraint: it forces the agency to build an evidentiary record, to respond to the evidence submitted, and to defend the final findings against that record in court. The process is slow partly because it is adversarial and evidence-bound, and that is a feature of law rather than a failure of it.

Treaty obligations add a second genuine constraint, and this one runs in the opposite direction from the politics-only story. Where control is required by United States obligations under international treaties, the statute commands the Attorney General to impose it without the usual findings and without the usual procedure, and a separate procedure governs the American response to international scheduling decisions under the psychotropic convention, including publication in the Federal Register and opportunity for public comment. The treaty machinery turns because the international obligation requires it to, and the domestic analysis of that constraint, with its history in the marijuana proceedings, is set out in the section above. Whatever one thinks of the underlying drug policy, the mechanism is not discretionary theater; it is a statutory command triggered by an international legal obligation, with the Single Convention on Narcotic Drugs and the Convention on Psychotropic Substances sitting behind it.

The record of actual movement supplies the third answer. Substances do move down and off the schedules, which is what a purely political system would have no reason to permit. Epidiolex, a purified cannabidiol oral solution, was placed in Schedule V following its 2018 approval for seizure disorders, and the DEA confirmed in March 2020 that it was no longer subject to the Controlled Substances Act at all, removing the only CBD medicine approved by the Food and Drug Administration from control entirely. Hydrocodone combination products traveled the other way, moving from Schedule III to Schedule II under a final rule that took effect on October 6, 2014, after a scientific and medical evaluation concluded that the combination products carried a high potential for abuse. The pattern across these episodes is consistent with the statute’s design rather than with politics alone: upward movement and treaty-driven placement proceed through fast statutory lanes, while downward movement requires the slow road of formal rulemaking, a complete factor analysis, and an enforcement agency willing to write a rule that reduces its own authority. The honest account must also concede what the veto cannot explain. Some placements persist because the science genuinely has not settled, some because the enforcement apparatus has institutional reasons to prefer the status quo, and some because no interested party has pressed a petition forcefully enough to make inaction politically costly. The statute supplies the one-way veto and the uneven lanes; the actors inside the system supply the rest.

That asymmetry of lanes also explains a detail that surprises many readers. The same statute that takes years to move a substance down a schedule lets the Attorney General place a substance in Schedule I temporarily, for up to two years with a possible one-year extension, on a finding that temporary control is necessary to avoid an imminent hazard to the public safety. Temporary scheduling considers only three of the eight factors, abuse history and pattern, scope and duration of abuse, and public health risk, and the resulting order is not subject to judicial review. The temporary lane was built for speed in one direction, just as the formal lane was built for deliberation in both. The fentanyl-related substances class received a temporary Schedule I order in February 2018, which held the structural class under control while the government worked through the question of permanent treatment.

A second fast lane runs through immediate precursors. The Attorney General may place an immediate precursor in the same schedule as the controlled substance it produces, or in any schedule with a higher numerical designation, without regard to the findings otherwise required and without regard to the ordinary procedures. The logic is containment: once a substance is controlled, the chemicals one step upstream from it can be pulled onto the schedules by order rather than by rulemaking. The statute even adds a limiting principle of its own, providing that substances may not be scheduled solely because they are precursors of a designated immediate precursor, which keeps the chain from extending indefinitely. And a third set of valves operates at retail level. The statute expressly provides that dextromethorphan, the common cough suppressant, is not deemed included in any schedule by reason of the statute’s enactment, a congressional carve-out written directly into the text. The same subsection lets the Attorney General exempt compounds, mixtures, and preparations containing controlled substances when they meet specified categories, such as prescription mixtures whose other ingredients are present in combinations that vitiate the abuse potential, or non-human-use preparations packaged so they present no significant abuse potential. These provisions show the statute at its most granular, adjusting control to the actual risk of the preparation rather than the bare identity of the ingredient.

A final note on reading the section numbers will help, because the two most-cited provisions are constantly confused. Section 811 is the machinery. It contains the rulemaking authority, the hearing requirement, the health agency evaluation, the eight factors, the treaty provisions, the immediate precursor rule, the abuse-potential referral, the over-the-counter exclusions, the temporary scheduling authority, and the interim final rule procedure. Section 812 is the inventory and its entry tickets. Subsection 812(a) establishes the five schedules, and subsection 812(b) sets the findings required for placement in each, such as the three Schedule I findings of high abuse potential, no currently accepted medical use in treatment in the United States, and lack of accepted safety under medical supervision. The habit that keeps a reader oriented is simple: if the passage describes a procedure, a petition, a hearing, or a finding, it belongs to 811; if it describes what a schedule means or which substances sit in which schedule, it belongs to 812. The inventory side has its own history worth knowing. The original schedules were not created by agency rulemaking at all; Congress placed the initial substances in the schedules directly, in subsection 812(c), when it enacted the Controlled Substances Act as title II of the Comprehensive Drug Abuse Prevention and Control Act of 1970. The section then handed the Attorney General the rulemaking machinery of 811 to keep the lists current, and it directed that the schedules be updated and republished, semiannually at first and annually thereafter. So when a reader sees a substance move schedules in a Federal Register notice, the authority behind the move is almost always 811; when a reader sees the schedules themselves, with heroin and marihuana on the Schedule I list as Congress wrote it, that is 812 speaking. Readers tracing a citation that stops at (h) are looking at the older map, as explained above. The series companion explainer on how to read a federal statute gives the full method for decoding statutory citations, cross-references, and amendment histories.

The honest account, then, is neither a conspiracy nor a laboratory. The statute imposes a genuine scientific evaluation, binds the enforcement agency to the health agency’s science in the negative direction only, subjects treaty-driven placements to a real international obligation, and moves substances in both directions through lanes of deliberately different speeds. The process has rules, and the rules have edges, and the edges are where the politics of drug policy grind against the science. Recognizing the veto without imagining a pen is the first step toward reading the scheduling record accurately.

The Schedule and Process Table

The full system compresses into a single reference. The first half of the table states the three statutory findings for each schedule with representative examples; the second half states the procedural steps of a scheduling proceeding with the actor, the required input, and the binding effect at each stage.

Element	What the statute requires	Examples or binding effect
Schedule I	High abuse potential; no currently accepted medical use; lack of accepted safety under medical supervision	Heroin, LSD, marijuana, peyote, methaqualone
Schedule II	High abuse potential; accepted medical use, or accepted medical use with severe restrictions; severe dependence	Cocaine, methamphetamine, morphine, oxycodone, fentanyl, Adderall, Ritalin
Schedule III	Abuse potential less than schedules I and II; accepted medical use; moderate or low physical or high psychological dependence	Codeine combination products, ketamine, anabolic steroids
Schedule IV	Low abuse potential relative to schedule III; accepted medical use; limited dependence relative to schedule III	Xanax, Valium, Ativan, Klonopin, Ambien, tramadol
Schedule V	Low abuse potential relative to schedule IV; accepted medical use; limited dependence relative to schedule IV	Robitussin AC, Lomotil, Lyrica
Initiation	Actor: Attorney General through the DEA; input: own motion, the Secretary’s request, or the petition of any interested party	Opens the proceeding; no placement changes yet
HHS evaluation	Actor: Secretary of HHS through the FDA; input: written scientific and medical evaluation under the eight factors	Advisory as to the schedule; a recommendation against control binds
Binding recommendation	Actor: the Secretary; input: a recommendation that the substance not be controlled	The Attorney General shall not control the substance
Rulemaking	Actor: the DEA; input: a substantial-evidence determination and Federal Register notice	On the record after opportunity for a hearing
Hearing	Actor: an administrative law judge; input: testimony and evidence from any interested person	Certified record and recommended decision
Final order	Actor: the DEA Administrator; input: the certified record and filed exceptions	Published in the Federal Register; effective in 30 days; reviewable under 21 U.S.C. 877

Reading the System as a Whole

The machinery this article has walked through can be held in a single paragraph. Three findings define each schedule, and the most restrictive schedule demands all three at once. The enforcement agency initiates on its own motion, at the Secretary’s request, or on the petition of any interested party. The health agency evaluates the science, and its recommendation against control ends the matter while its recommendation for a lower schedule does not compel action. Eight factors organize the evidence, an adversarial hearing tests it, and a reviewing court examines only the record the hearing produced. Temporary orders act fast and expire on their own clock, treaty obligations set a floor that domestic discretion cannot go below, and the documented movements in both directions show a system that revises its judgments when the evidence moves, even if the revision arrives slowly.

That paragraph is worth keeping in mind when reading the rest of the cluster, because each sibling article takes up one thread this one had to leave at the statutory surface. The provisions examined here live inside the Controlled Substances Act itself, and the guide to that statute shows how the scheduling machinery fits alongside the Act’s prohibitions, penalties, and regulatory apparatus. What accepted medical use actually requires in practice is a separate inquiry, and the article on FDA approval mechanics explains the evidentiary standard from which the scheduling standard borrows its central demand for adequate and well-controlled studies. The consequences of placement, from prescribing limits to the conflict between federal schedules and state law, are developed in the article on federal drug law and state law, which takes up where a scheduling decision leaves off. The section-numbering note on subsections (a) through (j) is part of a larger decoding skill, and the companion explainer on how to read a federal statute gives the full method for tracing citations, cross-references, and amendment histories across any legislation. And the whole domain sits inside a longer legislative history, which the article on American health legislation since 1950 places in its proper sequence.

Readers working through this material for coursework or examination preparation will find the schedules and the procedure easier to retain with structured study habits. The series companion notebook supports that work through guided legislation study notes, and a second companion offers civics study drills and practice questions for testing recall of the schedules, the factors, and the procedural sequence. The statute rewards readers who learn it as a procedure rather than as a list: who acts, what input the action requires, and what binds the actor once the input arrives.

Frequently Asked Questions

Q: How does drug scheduling actually work?

Drug scheduling is the administrative process that places substances onto, moves them between, or removes them from the five schedules of the Controlled Substances Act. The proceeding can begin three ways: the Attorney General, acting through the Drug Enforcement Administration, may act on the agency’s own motion, at the request of the Secretary of Health and Human Services, or on the petition of any interested party. Before acting, the Attorney General must request a scientific and medical evaluation and recommendation from the Secretary, prepared through the Food and Drug Administration under eight statutory factors. If the Secretary recommends against control, the Attorney General may not control the substance. When the proceeding goes forward, the agency must make the three findings required for the destination schedule in 21 U.S.C. 812(b), conduct formal rulemaking on the record after opportunity for a hearing before an administrative law judge, and issue a final order supported by the hearing evidence. Any aggrieved person may seek judicial review, and the court sustains findings supported by substantial evidence.

Q: What are the five schedules in drug scheduling?

Schedule I holds substances with high abuse potential, no currently accepted medical use in treatment in the United States, and a lack of accepted safety under medical supervision, which is why nothing in Schedule I may be lawfully prescribed; heroin and LSD are the archetypes. Schedule II shares the high abuse potential finding but adds accepted medical use, so cocaine, morphine, oxycodone, fentanyl, and medicines such as Adderall and Ritalin sit there under tight restriction, with prescriptions that cannot be refilled. Schedule III requires abuse potential less than the substances in Schedules I and II, accepted medical use, and moderate or low physical dependence or high psychological dependence; codeine combination products and ketamine are members. Schedule IV requires low abuse potential relative to Schedule III and limited dependence relative to Schedule III; medicines such as Xanax, Valium, Ativan, and Ambien sit there. Schedule V is the floor, with low abuse potential relative to Schedule IV and limited dependence relative to Schedule IV, covering preparations such as Robitussin AC and Lyrica.

Q: Who decides drug scheduling, the DEA or HHS?

Both agencies decide, but they decide different halves of the question. The Attorney General, acting through the Drug Enforcement Administration, initiates proceedings and issues the final scheduling orders after formal rulemaking. The Secretary of Health and Human Services supplies the scientific and medical evaluation and recommendation that the statute requires before proceedings begin, and that evaluation is prepared through the Food and Drug Administration with data from the National Institute on Drug Abuse. The recommendations bind the Attorney General as to scientific and medical matters, and if the Secretary recommends that a substance not be controlled, the Attorney General shall not control it. The health agency therefore holds a one-directional veto over control, while the enforcement agency holds the pen for the final placement decision, including whether to act on a recommendation favoring a lower schedule or removal. No single office decides alone.

Q: Is the HHS recommendation binding in drug scheduling?

The recommendation is binding in one direction and advisory in the other. Under 21 U.S.C. 811(b), the Secretary’s recommendations to the Attorney General are binding as to scientific and medical matters, and if the Secretary recommends that a substance not be controlled, the Attorney General shall not control the substance. That negative recommendation ends the proceeding before it begins, and no amount of law enforcement evidence can lawfully overcome it. The statute contains no mirror-image command. Nothing requires the Attorney General to place a substance in the schedule the Secretary recommends, or to act at all on a recommendation favoring a lower schedule or removal, and the Office of Legal Counsel confirmed during the marijuana rescheduling proceedings that the health agency’s placement recommendation was not binding on the Drug Enforcement Administration. The health agency can forbid a result but cannot command one: a veto without a pen.

Q: What is emergency drug scheduling?

Emergency drug scheduling, called temporary scheduling in the statute, is the fast lane Congress added to 21 U.S.C. 811(h) through the Comprehensive Crime Control Act of 1984. When the Attorney General finds that temporary placement of a substance in Schedule I is necessary to avoid an imminent hazard to the public safety, the agency may act by order without the ordinary scientific and medical evaluation, considering only three of the eight factors: history and pattern of abuse, scope and significance of abuse, and public health risk. The substance must be unscheduled and unapproved, thirty days must pass after notice in the Federal Register and to the Secretary of Health and Human Services, and the order expires after two years, with one possible one-year extension while permanent proceedings are pending. The order is not subject to judicial review. The mechanism was built for waves of designer drugs, and its most ambitious use was the February 2018 class-wide order covering fentanyl-related substances.

Q: Do treaties limit drug scheduling?

Treaties limit how far control can be undone, and the constraint operates as a floor rather than a ceiling. Under 21 U.S.C. 811(d), when control of a substance is required by United States obligations under international treaties, conventions, or protocols in effect on October 27, 1970, the Attorney General shall issue an order controlling the substance under the schedule deemed most appropriate to carry out those obligations, without regard to the usual findings and procedures. The Single Convention on Narcotic Drugs of 1961 is the principal treaty at issue, and a domestic decision to remove a treaty-covered substance entirely would collide with the obligation to keep it under control. The floor does not dictate every classification above the minimum: the statute preserves the Attorney General’s judgment about what the obligations require, and nothing stops the United States from maintaining stricter controls than the convention demands. A separate procedure, added by the Psychotropic Substances Act of 1978, governs the American response to international decisions under the 1971 psychotropic convention, with Federal Register notice and public comment before the domestic machinery implements the result.

Q: Why is Schedule I circular in drug scheduling?

Rescheduling petitioners and research advocates, including NORML and MAPS, identify a loop between the evidentiary standard for accepted medical use and the restrictions that Schedule I placement imposes on research. The Drug Enforcement Administration’s 1992 five-part test requires adequate and well-controlled studies proving efficacy, borrowing the standard the Food and Drug Administration applies before approving a new drug. Conducting those studies on a Schedule I substance requires a DEA researcher registration and FDA protocol review, and for marijuana the supply was for decades restricted to a single grower under contract to the National Institute on Drug Abuse, with an additional Public Health Service review imposed from 1999 to 2015. The criticism, stated by NORML, MAPS, and allied researchers, is that the placement withholds the registrations, funding priority, and supply the studies require, and then the absence of those studies is cited to justify keeping the placement. The D.C. Circuit came closest to naming the problem in 1991, striking down three earlier agency factors as logically impossible for any Schedule I substance to satisfy, though it upheld the revised 1992 test in 1994.

Q: Has anything ever been moved in drug scheduling?

Substances have moved in both directions, and some have left the schedules entirely. Dronabinol, synthetic delta-9 tetrahydrocannabinol sold as Marinol, began in Schedule I, moved to Schedule II by a 1986 final rule, and reached Schedule III in 1999 after its marketed record supplied the evidence the statute asks for. Buprenorphine moved up from Schedule V to Schedule III by a 2002 final rule. Hydrocodone combination products traveled from Schedule III to Schedule II under a final rule effective October 6, 2014, driven by diversion and misuse data. Congress has also moved substances by statute: the Hillory J. Farias and Samantha Reid Date Rape Drug Prohibition Act of 2000 placed gamma hydroxybutyric acid in Schedule I while preserving the approved medicine Xyrem in Schedule III, and the Anabolic Steroids Control Act of 1990 placed steroids in Schedule III. Epidiolex, a purified cannabidiol medicine approved in 2018, was placed in Schedule V and then removed from control entirely in March 2020. Propoxyphene left the schedules after the FDA requested its market withdrawal in 2010.

Q: What are the eight factors under section 811(c)?

The statute commands the Attorney General to consider eight factors for every substance proposed for control or removal: actual or relative potential for abuse; scientific evidence of pharmacological effect, if known; the state of current scientific knowledge; history and current pattern of abuse; scope, duration, and significance of abuse; public health risk; psychic or physiological dependence liability; and whether the substance is an immediate precursor of an already controlled substance. The Department of Health and Human Services weighs factors two, three, six, seven, and eight in full, plus the scientific and medical dimensions of the rest, while the Drug Enforcement Administration develops the law enforcement evidence on abuse potential and abuse patterns. The published eight-factor analyses are posted to the public docket for each scheduling action, so manufacturers, medical societies, patient groups, and reviewing courts can test each factor’s treatment. The factors apply to the full life cycle: scheduling, rescheduling, and removal all pass through the same eight inquiries.

Q: Who can file a scheduling petition?

Any interested party may petition, a category the statute and the courts read broadly. Patients, physicians, manufacturers, trade groups, advocacy organizations, researchers, and ordinary citizens with a demonstrable interest have all entered the process this way. Under 21 CFR 1308.43, a petition must set out the proposed rule, state the petitioner’s interest, and include a reasonably concise statement of the supporting facts, including a summary of any relevant medical or scientific evidence known to the petitioner. The Administrator must notify the petitioner whether the petition is accepted for filing and explain any refusal. A petition that cannot support a finding of abuse potential or that fails to engage the eight statutory factors ends with a denial letter, and a denial that constitutes a final decision is reviewable in court. Petitions have driven major proceedings: a physician’s 1999 petition set in motion the hydrocodone transfer finalized in 2014, and NORML’s 1972 petition opened the decades-long cannabis proceedings.

Q: What limits apply to temporary scheduling under section 811(h)?

Temporary scheduling is fenced on every side to keep it temporary. Only the Attorney General may issue the order, acting through the DEA Administrator, and only on a finding that temporary Schedule I placement is necessary to avoid an imminent hazard to the public safety. The substance must not be listed in any other schedule and must have no exemption or approval in effect under section 505 of the Federal Food, Drug, and Cosmetic Act, so no approved medicine can be swept in. The Attorney General considers only three of the eight factors, the abuse-history and public-health evidence that can be gathered quickly, and must give thirty days’ notice in the Federal Register and to the Secretary of Health and Human Services before the order issues. The order expires at the end of two years, with one possible one-year extension only while permanent proceedings are pending, and it is vacated when those proceedings conclude. The order is not subject to judicial review, a tolerable exemption only because the control cannot endure.

Q: How do treaties constrain domestic scheduling under section 811(d)?

The constraint is a one-way floor. When United States obligations under a treaty in effect on October 27, 1970 require control of a substance, the Attorney General shall issue an order placing it in the schedule deemed most appropriate to carry out those obligations, skipping the usual findings and the usual procedures. The Single Convention on Narcotic Drugs of 1961 is the treaty most often at issue. In the marijuana proceedings, the Bureau of Narcotics and Dangerous Drugs refused even to accept a 1972 petition on treaty grounds, and the D.C. Circuit held in 1977 that treaty obligations prevented placement below Schedule II, a reading the agency repeated in its August 2016 denial. A later Office of Legal Counsel analysis concluded the convention did not require Schedule I or II placement, showing that the floor’s exact height is contested. For the 1971 psychotropic convention, a separate procedure requires Federal Register notice, public comment, and an American scientific evaluation before the domestic machinery implements an international decision.

Q: How does judicial review of a scheduling decision work?

Section 877 of title 21 gives any person aggrieved by a final scheduling decision the right to petition for review in the Court of Appeals for the District of Columbia or in the circuit where the person’s principal place of business is located, within thirty days after notice of the decision. Findings of fact by the Attorney General, if supported by substantial evidence, are conclusive on review. That deferential standard is why the hearing record matters so much: the court will not reweigh the evidence, so challengers must build their case inside the agency proceeding through testimony and exhibits rather than persuading a court from scratch. The standard explains the cannabis case law, where the D.C. Circuit repeatedly upheld denials of rescheduling petitions while policing the agency’s test, striking down three factors in 1991 as logically impossible for a Schedule I substance to satisfy and upholding the revised five-part test in 1994. Temporary scheduling orders are the exception: the statute makes them not subject to judicial review at all.

Q: What is dependence liability in drug scheduling?

Dependence liability is the seventh of the eight statutory factors, and the statute’s phrasing preserves a distinction that shapes the schedule definitions. Psychic dependence means craving, compulsive seeking, and continued use despite harm; physiological dependence means tolerance and the withdrawal syndrome that follows when use stops. The schedule findings calibrate the inquiry on a descending scale: Schedule II requires abuse that may lead to severe psychological or physical dependence, Schedule III to moderate or low physical dependence or high psychological dependence, and Schedules IV and V to limited dependence measured against the schedule above. When the Drug Enforcement Administration evaluates a newly approved medicine, animal and human studies of withdrawal and of continued drug-seeking after access is restricted supply much of the evidence, supplemented by postmarketing reports. A drug with a punishing withdrawal profile lands higher than a pharmacological cousin with a gentler one, even when the two share a therapeutic purpose.

Q: Why does section 811 run from subsection (a) to subsection (j)?

The section grew by amendment, and citations that stop at (h) are reading the original 1970 map. Subsection (h), the temporary scheduling authority, arrived through the Comprehensive Crime Control Act of 1984. Subsection (i), covering the temporary and permanent scheduling of anabolic steroids, arrived in 2014 through the Designer Anabolic Steroid Control Act. Subsection (j), the interim final rule procedure under which the Attorney General must control a newly approved drug with abuse potential within ninety days of the Secretary’s recommendation, arrived in November 2015. The distinction between the two sections matters more than the lettering: section 811 is the machinery, containing the rulemaking authority, the hearing requirement, the health agency evaluation, the eight factors, the treaty provisions, the precursor rule, and the temporary scheduling power, while section 812 is the inventory, with subsection 812(a) establishing the five schedules and 812(b) setting the findings required for each.

Q: What is an immediate precursor in drug scheduling?

An immediate precursor is a chemical that sits one step upstream from an already controlled substance and exists principally to be converted into it in clandestine laboratories, and the statute gives the government a fast lane against the supply chain. The eighth scheduling factor asks whether a substance is an immediate precursor of a controlled drug, and section 811(e) lets the Attorney General place an immediate precursor in the same schedule as the controlled substance it produces, or in any schedule with a higher numerical designation, without regard to the findings otherwise required and without regard to the ordinary procedures. The word immediate does real interpretive work: it excludes distant raw materials with broad legitimate industrial uses and reaches only the compounds whose principal destination is conversion. The statute adds its own limiting principle, providing that a substance may not be scheduled solely because it is a precursor of a designated immediate precursor, which keeps the chain from extending indefinitely upstream.

Q: Why are downward scheduling moves rare?

The statute supplies a one-way veto and lanes of deliberately different speeds. The Secretary’s recommendation against control binds the Attorney General, but nothing compels the Attorney General to act on a recommendation favoring a lower schedule or removal, so a favorable health agency finding arrives at a desk where it can be respectfully received and set aside. A downward move must travel the slow road of formal rulemaking on the record after opportunity for a hearing, with findings made against each destination schedule’s criteria, and it asks the enforcement apparatus to write a rule that reduces its own authority. The Office of Legal Counsel confirmed during the marijuana proceedings that the health agency’s placement recommendation did not bind the Drug Enforcement Administration. Upward movement and treaty-driven placement, by contrast, travel fast statutory lanes: temporary orders act in weeks, immediate precursors move by order, and Congress can place substances by legislation. The asymmetry is deliberate, and the tempo of the schedules follows it.

Q: Has any substance ever been descheduled entirely?

Removal from the schedules is rarer than transfer between them, but the record includes clean examples. Epidiolex, a purified cannabidiol oral solution approved by the Food and Drug Administration in 2018 for seizure disorders, was first placed in Schedule V, and the Drug Enforcement Administration confirmed in March 2020 that it was no longer subject to the Controlled Substances Act at all, removing the only approved CBD medicine from control entirely. Propoxyphene, marketed as Darvon and Darvocet in Schedule IV, followed its market exit off the schedules: after the FDA asked in November 2010 that propoxyphene products be withdrawn from the United States market because of fatal heart rhythm risks, the DEA removed the drug from control, ending the federal scheduling of a medicine with no lawful market left to regulate. Removal runs on the same eight factors in reverse, requiring a finding that the substance no longer meets the requirements for any schedule, and treaty obligations must also permit the removal.

Q: How does FDA approval relate to accepted medical use?

The scheduling standard for accepted medical use borrows its evidentiary core from the drug approval standard. The Drug Enforcement Administration’s five-part test for currently accepted medical use, set out in a March 1992 final rule and upheld by the D.C. Circuit in 1994, centers on adequate and well-controlled studies proving efficacy, which is the standard the Food and Drug Administration applies before approving a new drug. An approved medicine therefore carries the clinical trial evidence the scheduling test demands, which is why approved medicines so rarely sit in Schedule I and why Marinol moved from Schedule I to Schedule II to Schedule III only after being studied and marketed as a pharmaceutical. Approval is not the same as a scheduling finding: the FDA judges safety and efficacy for a proposed use, while the scheduling findings weigh abuse potential, dependence, and the comparison against the substances already sorted into neighboring schedules. A new drug with abuse potential is forwarded to the Attorney General at approval, and a health agency recommendation for control triggers an interim final rule within ninety days.

Q: What happens at a formal scheduling hearing?

Scheduling rules are made on the record after opportunity for a hearing under sections 556 and 557 of the Administrative Procedure Act, not through ordinary notice-and-comment rulemaking. After the Federal Register notice, any interested person may request a hearing or waive the opportunity with a written statement of position, and at least thirty days must pass before the hearing begins. An administrative law judge hears testimony, receives evidence subject to rules of relevance and materiality, with opportunity for cross-examination, then certifies the record and issues a recommended decision. Participants may file written exceptions, and the Administrator considers them with the certified record before issuing the final order, which may adopt, modify, or reject the judge’s recommendation. The order publishes in the Federal Register and takes effect thirty days later. Because the final order must rest on the hearing evidence, and because courts review findings only for substantial evidence, the hearing is the main engine of fact-finding and the only record that will ever be reviewed.