A statute on paper and a system in operation

USDA food safety begins on paper, in statutory definitions enacted decades apart, and ends in operation, in whether an inspector stands on a slaughter line during every operating hour, and whether an agency can order contaminated product removed from commerce or can only ask the establishment that made the product to remove it. The distance between those two points is large. Enacted text supplies tidy categories, naming meat, poultry, and egg products on one side and assigning the remainder of the food supply to a general food statute on the other side. Operating reality supplies establishments that make both kinds of product under one roof, recipes that cross a jurisdictional threshold by a few percentage points, and hazards that no examination of a carcass can detect.

The two agency split is the first expression of that gap. The Food Safety and Inspection Service, inside the Department of Agriculture, administers the inspection statutes for meat, poultry, and processed egg products. The Food and Drug Administration, inside the Department of Health and Human Services, administers the Federal Food, Drug, and Cosmetic Act for most other food. The Government Accountability Office, in report GAO-17-74 issued in 2016, described the Food and Drug Administration as responsible for roughly four fifths of the United States food supply and the inspection service as responsible for roughly one fifth, while the two agencies received close to the same food safety funding. The disparity in resources follows from a disparity in legal duty. The smaller share carries a statutory obligation of continuous presence in slaughter establishments, with examination of animals before slaughter and of every carcass and its parts after slaughter, while the larger share is overseen through periodic, risk based inspection of facilities.

The statutory inheritance behind the split explains its shape. Congress enacted the Federal Meat Inspection Act on June 30, 1906, the Poultry Products Inspection Act on August 28, 1957, and the Egg Products Inspection Act on December 29, 1970, each statute placing a defined product class inside the Agriculture Department inspection system. Shell eggs remained under the Food and Drug Administration, while eggs removed from the shell and processed as liquid, frozen, or dried products moved to the inspection service. Siluriformes fish, including catfish, moved to the inspection service under the farm act of 2008, while all other seafood remained with the Food and Drug Administration. Each assignment rested on the commodity Congress addressed at the moment of enactment, and endured through later amendment, so that jurisdiction came to follow ingredient, form, and statutory history more faithfully than hazard.

The enforcement constraint at the center of the system has a narrower source, a single contested word in a statutory definition. After an outbreak of Escherichia coli O157:H7 infection traced to undercooked ground beef in January 1993, in which the Centers for Disease Control and Prevention retrospective account counted more than 700 illnesses across four states and four children killed, the inspection service treated that organism as an adulterant in raw ground beef, and the courts sustained that treatment because ordinary cooking practice, as consumers actually practiced cooking, did not reliably destroy the organism. Salmonella stood in the opposite position. In Supreme Beef Processors, Inc. v. United States Department of Agriculture, decided by the Fifth Circuit in December 2001, failure of a Salmonella performance standard could not support withdrawal of inspection, because the organism occurs naturally in raw meat and poultry and normal cooking destroys it. One pathogen could be condemned through the definition of adulteration. The other could be measured, published, and pursued only through sanitation and hazard system failures.

The remaining pieces of the operating system follow from those two foundations, inheritance and definition. Rulemaking, label approval before sale, detention of product in commerce for up to 20 days, seizure through a court, and suspension of inspection supplied substantial tools. A mandatory recall order was never among those tools, although repeated bills proposing that authority were not enacted through February 1, 2021, and proposals to merge the divided structure into a single food safety agency met the same outcome. The system that resulted is an inheritance administered daily, bounded at its most consequential point by judicial construction of the language Congress left in place.

The two agencies and the statutes that divide the food supply

Federal food safety in the United States begins with a formal division of legal identity, and the division has to be stated precisely before any question of practice can be understood. The Food Safety and Inspection Service is a component of the Department of Agriculture. The Food and Drug Administration is a component of the Department of Health and Human Services. Neither agency holds a general commission over everything Americans eat. Each holds the authority that particular statutes confer, and the statutes confer authority by naming classes of products rather than by describing food as a single field. The inspection service administers inspection laws written for meat, for poultry, and for processed egg products. The Food and Drug Administration administers a broader food law written for the remainder of the supply. The result is a system in which jurisdiction follows the product, departmental placement follows the statute, and the statute follows decisions Congress made at very different moments for very different reasons. Any account that treats the two agencies as interchangeable regulators with overlapping preferences misses the legal structure. Authority in this field is not assigned by expertise in the abstract. Authority is assigned by enacted text, and enacted text in this field arrived product by product.

That structure explains why the names of the agencies can mislead a reader who expects symmetry. The Food Safety and Inspection Service sounds, by its title, like the federal government’s general food safety agency, and the Food and Drug Administration sounds, by its title, like an agency concerned principally with medicines. The statutory record supports neither impression in a simple form. The inspection service is the narrower of the two in product reach, confined by its governing acts to defined categories of animal products and to the establishments that slaughter, process, and handle those products under federal inspection. The Food and Drug Administration, despite a name that places drugs beside food, holds responsibility for the larger share of the food supply under a statute that reaches food as a general category. The asymmetry is therefore built into the founding assignments. One agency was built around a production process, the slaughter and processing of animals for food, and around the continuous presence that process was thought to require.

The oldest of the inspection statutes, and the foundation of the Agriculture Department side of the line, is the Federal Meat Inspection Act. Congress enacted the act on June 30, 1906, as part of the Meat Inspection Act, and the statute is codified at title 21, section 601 and following sections of the United States Code. The act addressed livestock slaughter and meat processing as activities that federal law would surround with mandatory inspection, and its central commands concerned the examination of animals before slaughter and the examination of carcasses and their parts after slaughter. The statute did not ask whether a particular establishment seemed risky and did not leave inspection to the discretion of the Secretary of Agriculture as a general supervisory tool. The statute made inspection a condition of lawful operation for covered establishments handling covered species. Congress returned to the statute in the Wholesome Meat Act of 1967, which substantially amended the 1906 act more than six decades after enactment, and the amended act remained the governing text for meat. The inspection service inherited a statute whose obligations were categorical, whose trigger was the identity of the product and the species behind the product, and whose enforcement logic ran through the presence of inspectors in establishments rather than through periodic visits alone.

Poultry received a separate statute rather than a place inside the meat act, and that legislative choice is itself part of the explanation for the fragmented structure. Congress enacted the Poultry Products Inspection Act on August 28, 1957, and the statute is codified at title 21, section 451 and following sections of the United States Code. The poultry act extended the inspection model to poultry and poultry products through a distinct legal instrument with its own definitions, its own coverage provisions, and its own administrative history. Congress strengthened the statute in the Wholesome Poultry Products Act of 1968, a year after the comparable amendment of the meat statute, and the pairing of the two amendments shows Congress working within the inherited structure rather than replacing that structure with a unified code. A legislator confronting poultry inspection in the middle of the twentieth century did not draft on a blank page. A meat inspection establishment system already existed inside the Department of Agriculture, with inspectors, establishments, marks of inspection, and a body of practice, and poultry legislation was fitted to that existing system through a parallel act. Parallelism, rather than consolidation, is the recurring legislative habit in this field, and the habit preserved departmental placement even as coverage expanded.

Egg products arrived later and by the same method. Congress enacted the Egg Products Inspection Act on December 29, 1970, and the statute is codified at title 21, section 1031 and following sections of the United States Code. The egg products act brought the processing of eggs removed from the shell, in liquid, frozen, and dried form, inside the Agriculture Department inspection system, with continuous inspection of egg products plants. The statute also fixed, by implication, one of the most instructive boundaries in the entire field, because shell eggs and egg products are the same underlying food at different stages of handling and yet fall on different sides of the agency line. Shell eggs in production, transport, storage, and retail fall to the Food and Drug Administration, while processed egg products fall to the inspection service, and the Department of Agriculture’s Agricultural Marketing Service grades shell eggs for quality only and not for safety. The egg legislation therefore did more than add a product class. The legislation demonstrated that the dividing principle is not the biological identity of a food, not the farm from which a food came, and not the nutritional character of a food. The dividing principle is the form in which a food is presented and the statute under which Congress placed that form.

Against those three product statutes stands the statutory base of the Food and Drug Administration. The Federal Food, Drug, and Cosmetic Act of 1938 supplies the foundation of the agency’s food authority, and Congress amended that act over the following decades, including by the FDA Food Safety Modernization Act, Public Law 111-353, signed on January 4, 2011. The contrast in statutory architecture is fundamental. The inspection acts name animal product categories and build a regime of establishment inspection around those categories. The 1938 act addresses food within a wider statute that also addresses drugs and cosmetics, and the food provisions operate across the general run of foods that no inspection act claims. Where an inspection act applies, the Agriculture Department agency governs. Where no inspection act applies, the 1938 act supplies the default law of the food supply, and the default covers an enormous range of production, from fresh produce and seafood other than one transferred fish category to packaged foods, beverages, and ingredients that contain too little meat or poultry to enter the inspection system. The default statute made the agency responsible for most of what the food system produces, while the older and more intensive inspection model was reserved for the categories Congress had singled out in 1906 and in the decades that followed.

Why are there two agencies for food safety in the United States?

Federal food safety is divided because Congress assigned meat, poultry, and processed egg products to the Department of Agriculture under inspection statutes enacted in the early twentieth century, and assigned most other foods to the Food and Drug Administration under a separate statute, and neither assignment was later consolidated into a single agency.

The honest explanation for that arrangement is historical sequence rather than institutional design. No Congress sat down with a complete map of the food supply and apportioned the map between two departments according to a theory of regulatory competence. The meat inspection act of 1906 belonged to the slaughterhouse era, when the federal problem that demanded a statutory answer was the condition of animals and carcasses in large slaughtering and packing establishments and the confidence of domestic consumers and foreign buyers in American meat. The institutional answer was placed where livestock and slaughter already belonged in the federal government, inside the Department of Agriculture, and the answer took the form of inspectors stationed in establishments. Once that answer existed in law, later Congresses extended the pattern by analogy. Poultry followed meat because poultry slaughter resembled meat slaughter. Egg products followed because breaking and processing eggs resembled other inspected processing. Each extension was reasonable when viewed from the statute that preceded the extension, and the accumulation of reasonable extensions produced a structure that no single enactment would have designed. Product classes were frozen into statutes at the moment each class became a federal concern, and the statutes then governed long after production, distribution, and retailing had changed around them.

Amendment practice reinforced the freeze. When Congress modernized meat inspection in 1967 and poultry inspection in 1968, the amendments strengthened and revised the separate acts rather than merging the acts into a single animal products code or moving the programs to the agency holding the 1938 act. When Congress enacted a major food safety modernization statute in 2011, that statute amended the Federal Food, Drug, and Cosmetic Act and rebuilt much of the Food and Drug Administration’s preventive framework, while the inspection acts on the Agriculture Department side continued under their own titles and their own logic. Two reform tracks therefore ran in parallel across more than a century, each track improving the statute in front of the legislators and neither track reconsidering the boundary between the tracks. Path dependence of that kind is not mere inertia. A mandatory inspection system creates establishments organized around inspection, a workforce of inspectors and veterinarians assigned to establishments, state programs modeled on the federal system, and congressional committee relationships organized around the department that administers the programs. Moving a program after such structures form means displacing far more than a line on an organization chart, and the costs of displacement help explain why a division born of sequence acquired the durability of a division born of principle.

The scale of the resulting division deserves exact statement, because casual description usually inverts the relationship between coverage and resources. The Government Accountability Office, in report GAO-17-74 issued in 2016, described the Food and Drug Administration as responsible for roughly 80 percent of the United States food supply and the Food Safety and Inspection Service as responsible for roughly 20 percent, or in round terms roughly four fifths of the supply on one side of the line and one fifth on the other. The same Government Accountability Office formulation placed the food safety funding of the two agencies close to the same level, despite the four to one disparity in shares of the supply. The disparity is not, without more, proof that either agency holds too much or too little money, because the inspection service carries a statutory obligation of continuous presence in slaughter establishments that a periodic inspection regime does not carry, and presence is expensive in personnel. The disparity does establish the central fiscal fact of the divided system. The smaller product share consumes resources comparable to the larger share because the older statute purchases a different and more labor intensive kind of oversight, and Congress funds the obligation the statute creates rather than funding food safety in proportion to the volume or variety of food each agency oversees.

Personnel figures from an earlier period give that obligation a concrete shape, provided the figures are read with their date attached. A Congressional Research Service issue brief from 2003, IB10082, counted about 7,700 Food Safety and Inspection Service field employees, roughly 1,000 of whom were veterinarians, working at some 6,200 plants and import stations. Those counts describe the system as the brief recorded the system in that period and are not a census of the system in any later year. Even as dated figures, the counts illuminate the statutory design. A regime that conditions slaughter on the presence of an inspector, and that extends inspection across thousands of establishments, is a regime whose principal asset and principal cost is a deployed workforce. The Food and Drug Administration model, built on risk based and periodic inspection of facilities under statutory frequencies, deploys personnel in a different pattern and at a different rhythm.

The operative concept that connects the inspection statutes to individual animals and products is amenability. A species is amenable, in the usage of the inspection system, when one of the inspection acts subjects that species and products derived from that species to mandatory federal inspection. Amenability is a legal status and not a biological judgment, a judgment about danger, or a judgment about how a food is eaten. A species inside the statutory definitions enters a regime in which slaughter and processing occur under inspection as a condition of lawful commerce in the resulting products. A species outside those definitions does not enter that regime, however similar the species may be to a covered species in the kitchen or in the marketplace. Congress can change the status of a species by statute. Siluriformes fish, the order that includes catfish, were transferred to the inspection service by the farm act of 2008, and the transfer was implemented through a Food Safety and Inspection Service final rule in 2015 with inspection beginning in 2016, while the Food and Drug Administration retained responsibility for all other seafood. A single statutory transfer moved one fish category across the entire federal food safety structure, from one department to another, from one inspection philosophy to another, and the episode shows that the boundary, for all its durability, remains a legislative artifact that legislation can relocate.

The converse of amenability defines a substantial territory on the other side of the line. Exotic and game meats that are not amenable to the inspection acts fall to the Food and Drug Administration under the general food statute, so animals raised, hunted, or marketed outside the species coverage of the inspection acts are regulated without the mandatory establishment inspection that amenable species receive. The inspection service also offers voluntary inspection on a fee for service basis for some non amenable species, under which an establishment may purchase examination that the statute does not compel. The distinction between amenable and non amenable species therefore carries consequences that run from the staffing of establishments to the price of inspection itself. Mandatory inspection under the inspection acts is supported largely by congressional appropriations, reflecting a legislative judgment that the inspection of amenable species is a public condition of commerce. Voluntary inspection shifts the cost to the establishment that seeks the service. Two animals that reach a consumer in similar form can thus stand in entirely different legal relationships to the federal government, one inspected because a statute names the species, the other inspected only because a producer paid for inspection or else regulated under an entirely different statute in a different department.

Federal inspection is not the only inspection tier the statutes contemplate, because the inspection acts also accommodate state programs that operate inside the federal framework. State inspection programs meeting the federal standard, described in the statutory phrase as programs at least equal to the federal program, operate in roughly half the states and cover product handled within the state that maintains the program. The federal government does not surrender the standard in such an arrangement. A participating state enforces requirements equal to or stronger than the federal requirements under cooperative arrangements with the Department of Agriculture, and product inspected under a state program is, in the general structure of the system, confined to commerce inside that state. Alongside the intrastate tier, the Cooperative Interstate Shipment program, established by the farm act of 2008, permits selected establishments inspected under state programs to ship product across state lines, breaking the older equation under which state inspection meant intrastate commerce only. Selected establishments in a participating state program operate under the program with federal coordination, and product from those establishments may move in interstate commerce in a manner that state inspected product generally may not. The program is limited by design and selective in admission, and the limits themselves reveal the premise of the whole state structure. Interstate movement is the historical justification for the federal inspection presence, and every exception to the intrastate confinement of state inspected product has required a deliberate statutory decision.

Imports test the divided system at the border, and each side of the system applies a different theory of foreign control. The inspection service does not admit meat, poultry, or egg products from a foreign system on the strength of inspection applied shipment by shipment at the port. The governing theory is equivalence. A foreign inspection system must be determined equivalent to the United States system, with laws, controls, and inspection performance that achieve the same level of protection, before product from that system becomes eligible for entry, and imported product is then re inspected at import establishments inside the United States. The Food and Drug Administration applies a different border regime to the far larger range of foods under the 1938 act. Foreign food enters under a system built around information and importer responsibility, including prior notice of imported food shipments and, after the FDA Food Safety Modernization Act, foreign supplier verification programs under which importers must verify that foreign suppliers produce food in a manner that meets applicable United States standards. The contrast mirrors the domestic contrast. The inspection service asks whether a foreign government runs a system equivalent to the federal system and stations federal personnel to re examine what arrives. The Food and Drug Administration asks whether notice has been given and whether the importer has verified the supplier, placing legal responsibility on the private party that introduces the food into United States commerce. A shipment of beef and a shipment of produce may arrive through the same port on the same day, and the two shipments encounter different departments, different legal theories, and different distributions of responsibility between government and importer.

Departmental placement shapes more than border practice, because placement determines the congressional and budgetary environment in which each agency works. The inspection service sits inside a department whose historic constituency is agricultural production, and the service draws its funding largely from congressional appropriations considered within that department’s budget. The Food and Drug Administration sits inside a department organized around health and human services, and the agency’s food program competes for attention and resources inside an institution whose responsibilities extend well beyond food. Placement also determines oversight relationships. Agriculture committees in Congress guard the jurisdiction that the Department of Agriculture programs represent, and any proposal to move food safety functions out of the department encounters the institutional interests that placement creates. The divided system is therefore sustained by more than statutory text. Text assigns the products, appropriations sustain the workforce the text requires, and committee jurisdiction protects the assignment against structural revision, so that a division enacted for the conditions of one century persists as the working constitution of food safety in the next.

None of these abstractions, the statutes, the percentages, the programs, or the border theories, settles the question that a producer, a retailer, or a consumer actually faces, which is where the line falls on a real product sitting in a real plant or on a real shelf. A category name in a statute has to be applied to foods that combine ingredients, change form during processing, and cross the boundary between animal product and general food in the course of ordinary manufacture. The governing acts name meat, poultry, and egg products, and the general statute claims everything else, and between those formulations lies the daily work of deciding which department inspects a sandwich, a frozen meal, a soup, or a pizza, and which set of legal obligations follows from the decision.

Where the line falls: jurisdiction on the dinner plate

A dinner plate gives no reliable clue about which federal agency governed the food placed on it. Two items that sit side by side, look alike, travel through the same distribution system, and reach the same kitchen under similar conditions may have arrived there under entirely different statutes, different oversight regimes, and different enforcement tools. One item may have been produced in a plant visited every operating day by a federal inspector whose presence was a condition of lawful operation. Another item, separated from the first by a single ingredient or by the removal of a shell, may have been produced in a facility inspected only periodically on a risk-based schedule. The difference does not announce itself in taste, appearance, or price. It announces itself in law, in the definitions Congress wrote into separate statutes across separate decades, and in the agency practice that grew up around those definitions as new products tested old categories.

The plate is therefore a useful place to see federal food safety oversight in miniature. The larger division, described by the Government Accountability Office in GAO-17-74 in 2016, assigned roughly four fifths of the United States food supply to the Food and Drug Administration and roughly one fifth to the Food Safety and Inspection Service, even though the two sides received close to the same food safety funding. No shopper sorting a freezer case by hazard would draw the boundary where Congress and the agencies have drawn it, because the boundary was not drawn by sorting foods by hazard at all.

Why does a pepperoni pizza fall under a different inspector than a cheese pizza?

Frozen cheese pizza falls under the Food and Drug Administration, while frozen pepperoni pizza falls under the Food Safety and Inspection Service, because the pepperoni brings the product within the meat inspection statutes. The determining factor is ingredient and product category under those statutes, not a comparative judgment about which pizza presents greater danger.

The pizza example endures in discussions of federal food safety precisely because the contrast is so plain and so difficult to explain on grounds of risk alone. A frozen cheese pizza, consisting of crust, sauce, and cheese, is a food regulated by the Food and Drug Administration under the Federal Food, Drug, and Cosmetic Act of 1938, as amended. Add pepperoni, a meat topping prepared from inspected meat, in an amount sufficient to bring the finished product within agency practice for meat products, and the finished pizza becomes a product of the inspection service under the Federal Meat Inspection Act. The crust has not changed in character. The sauce has not changed. The cheese has not changed. What has changed is the legal category of the finished article, and with that category changes the plant oversight, the label review, and the enforcement relationship that stand behind the box in the freezer case.

The same plant may stand on both sides of that change. Both kinds of pizza can occur in the same establishment under different agency inspection. A facility producing frozen cheese pizzas and frozen pepperoni pizzas does not simply choose one regulator for the building and apply that choice to every line inside it. Instead, product by product and line by line, jurisdiction follows the article being made. One production run answers to one agency regime, and another run answers to the other. Plant management must therefore maintain, inside a single roof, working knowledge of two systems that differ in how often an inspector appears, how a label reaches the market, and what happens when a problem requires product to be pulled back from commerce.

The sandwich examples carry the same lesson into an even narrower space, where the arrangement of bread decides the regulator. A closed-faced meat sandwich, built with two slices of bread enclosing the meat, falls under the Food and Drug Administration. An open-faced meat sandwich, presented without the enclosing top slice, falls under the Food Safety and Inspection Service. That allocation, stated in congressional testimony quoted in Government Accountability Office era coverage of the split system, reflects longstanding jurisdictional practice rather than a finding that an open-faced sandwich exposes a diner to a categorically different danger. The meat in the two sandwiches may come from the same inspected source. The bread may come from the same bakery. The handling may be identical. The line turns on the form of the finished product as the agencies, working from statutory definitions and accumulated practice, have classified that form.

Inspectors, plant operators, and label reviewers need a rule that can be applied on the production floor without a fresh scientific inquiry into the relative risk of every variant of every product. A categorical line, even a line that produces odd neighbors, gives that administrability. A plant can be told, in advance and without litigation over each recipe, which agency will examine its operation and approve its labels. The cost of that clarity is a boundary that tracks form and ingredient history more faithfully than it tracks danger.

Eggs trace the line through a transformation rather than through an added ingredient. Shell eggs, eggs sold and handled in the shell, fall under the Food and Drug Administration in production, transport, storage, and retail. Liquid egg products, frozen egg products, and dried egg products, meaning eggs removed from the shell and processed into those forms, fall under the Food Safety and Inspection Service, with continuous oversight of egg products plants under the Egg Products Inspection Act of 1970. The biological material is the same at the start. An egg that remains in its shell travels under one regime, while the contents of an egg broken out of its shell, pooled, pasteurized, and packed as a liquid, frozen, or dried product travel under another. The act of breaking the shell and entering processing changes the regulator, the plant regime, and the relationship between the establishment and the federal presence inside it.

A third federal presence complicates the egg picture without displacing either safety regulator. The Agricultural Marketing Service, another component of the Department of Agriculture, grades shell eggs for quality only. A grade mark on a carton of shell eggs therefore should not be read as a safety determination by the Department of Agriculture, and the existence of a department grading program for shell eggs does not transfer safety jurisdiction over shell eggs away from the Food and Drug Administration. Three distinct functions meet at the same carton and the same breaking plant: quality grading by the Agricultural Marketing Service, safety oversight of shell eggs by the Food and Drug Administration, and safety oversight of processed egg products by the inspection service.

Seafood carries the line in the opposite direction, from a broad assignment to one agency toward a single carved out exception for the other. The Food and Drug Administration covers seafood generally. Catfish, meaning fish of the order Siluriformes, form the exception. The farm act of 2008 moved Siluriformes fish to the Food Safety and Inspection Service, and an inspection service final rule in 2015 implemented that transfer, with inspection beginning in 2016. All other seafood remained with the Food and Drug Administration. A fillet of catfish and a fillet of another fish may be processed, frozen, shipped, and cooked in closely comparable ways, yet the catfish enters the regime of the inspection service while the other fillet enters the facility regime of the Food and Drug Administration. The transfer was a legislative choice to place one defined group of fish inside a statutory program built originally for livestock and poultry.

Exotic and game meats show where the inspection statutes simply run out. Species amenable to the Federal Meat Inspection Act and the Poultry Products Inspection Act receive mandatory federal oversight within those programs. Meat from exotic and game species not amenable to those inspection acts falls to the Food and Drug Administration under its general food authority. For some non-amenable species, the inspection service offers voluntary fee-for-service examination, paid for by the establishment seeking it, rather than the mandatory program that applies to amenable species. A plant handling an amenable species cannot lawfully operate covered slaughter without the required federal presence. A processor handling a non-amenable species operates under the Food and Drug Administration regime unless a voluntary arrangement brings inspection service personnel into a defined fee-for-service role. The animal on the table does not decide the question by its size, its habitat, or the care taken in its preparation. Amenability, a statutory concept, decides it.

Between the clear cases lies a threshold that agency practice uses to sort mixed products containing small amounts of meat or poultry. Products containing more than 3 percent raw meat, or more than 2 percent cooked meat or poultry, are treated in inspection service practice as products of that service. Products at or below those small amounts may remain with the Food and Drug Administration even though meat or poultry is present as an ingredient. The precision matters, and so does the source of the line. Those percentages are stated as agency practice and regulation for sorting jurisdiction, not as a sentence written into the original meat inspection statute as a universal dividing rule for every mixed food. A soup, a sauce, a frozen meal, or a baked item can therefore change regulators when a recipe crosses a threshold measured in single percentage points. Formulation work that looks, from a culinary perspective, like a minor adjustment can become, from a regulatory perspective, a move from one federal system to another, with a new label pathway and a new oversight rhythm attached to the finished product.

Plants producing both kinds of product make the abstract boundary concrete in staffing, scheduling, and floor space. An establishment that makes a Food and Drug Administration product on one line and an inspection service product on another may host personnel from both agencies, each acting under a separate statute and a separate set of procedures. Records must satisfy two expectations. Labels must travel two different routes. A sanitation concern, a formulation change, or a records question may need to be presented in different terms depending on which side of the internal line the affected product occupies. Dual jurisdiction of that kind does not mean double oversight of the same article. Each article has its regulator. It means a single business, sometimes a single building and a single workforce, must operate fluently inside two legal systems at once, and must know, for every recipe and every run, which system governs before production begins rather than after product has entered commerce.

The explanation for lines of that kind lies in statutory inheritance. The Federal Meat Inspection Act dates to 1906, enacted on June 30 of that year as part of the Meat Inspection Act and substantially amended by the Wholesome Meat Act of 1967. The Poultry Products Inspection Act followed in 1957, enacted on August 28 of that year and strengthened by the Wholesome Poultry Products Act of 1968. The Egg Products Inspection Act arrived in 1970, enacted on December 29 of that year. The Food and Drug Administration built its food program under the Federal Food, Drug, and Cosmetic Act of 1938, later amended, including by the FDA Food Safety Modernization Act, signed on January 4, 2011. Each statute defined its field by the commodities and products Congress placed inside it at the time of enactment and amendment. Later Congresses added poultry as a separate program, added processed eggs as a separate program, and added catfish by amendment in 2008, without redrawing the entire map of federal food oversight around a single ranked list of hazards.

A risk ranking system, built from the ground up, would be expected to group foods by the likelihood and severity of harm they can carry, by the vulnerability of the populations eating them, and by the points in production where intervention does the most good. The inherited system groups foods, first, by whether a statute names the commodity, by whether a product falls inside a definition carried forward through amendment, and by whether agency practice has assigned a mixed or borderline article to one side of a definition. Pathogens do not respect those groupings. A hazard that can appear in meat can also appear in produce, in dairy, in seafood, or in a prepared food whose meat content sits below the jurisdictional threshold. The inheritance explains the map without proving that the map matches the distribution of danger, and much of the long debate over fragmentation, including Government Accountability Office placement of federal food safety oversight on its High-Risk List in 2007, has proceeded from that gap between inherited category and measured risk.

Statutory inheritance also explains why change tends to arrive commodity by commodity rather than as a general redrawing. When Congress moved catfish, Congress amended the assignment for Siluriformes and left the rest of seafood where the inherited structure had placed it. When Congress enacted the egg products statute, Congress addressed eggs removed from the shell and processed, and left shell eggs under the other agency. Accumulated across more than a century, those steps produced a plate on which neighboring items answer to different programs, not because any single decision maker compared every item on the plate and ranked it, but because no single decision maker ever held the whole plate in view with authority to rearrange it at once.

The practical consequences begin with time, specifically with how often a federal inspector is present while food is being made. Under the meat inspection statute, slaughter establishments operate with inspectors on the line during all operating hours, with antemortem examination of livestock and postmortem examination of every carcass and its parts, and an establishment cannot lawfully conduct covered slaughter without that coverage. Processing establishments under inspection service jurisdiction receive a federal presence at least once per operating day in the standard model. Egg products plants operate under continuous oversight in the inspection service model described for that program. A product that crosses into inspection service territory therefore enters a world in which federal presence is woven into the operating day itself, where absence of coverage can stop covered work, and where the relationship between plant and inspector is daily, direct, and structured by statute.

Across the line, the Food and Drug Administration model for food facilities is periodic and risk-based. The FDA Food Safety Modernization Act set statutory frequencies under which high-risk domestic facilities are inspected at least once every three years and other facilities at least once every five years. One model places the federal examiner inside the establishment as a recurring condition of production, above all in slaughter, where examination attaches to animals and carcasses. The other model allocates visits across a far larger universe of facilities according to assessed risk, with enforcement and follow up occurring between visits through other tools. A pepperoni pizza and a cheese pizza leaving the same plant therefore carry different temporal relationships to federal oversight: one was made under a regime of daily presence, and the other was made under a regime in which years may separate routine facility visits.

Labeling carries a second consequence that follows a product across the line just as faithfully. The inspection service operates a prior approval system for labels of meat, poultry, and egg products, expanded over time through generic approval for categories of labels, under which covered labels reach the market through a federal review pathway established before sale. Food labeling under the Food and Drug Administration is largely post-market, meaning that federal correction typically follows placement in commerce rather than preceding it through a comparable prior approval gate for ordinary food labels. A recipe change that moves a mixed product across the meat threshold can therefore change not only who visits the plant and how often, but also when the federal government sees the label and by what mechanism an error on that label is caught.

Recall authority carries a third consequence. The inspection service has no mandatory recall authority for meat, poultry, or egg products. A recall of such a product is the establishment action, undertaken voluntarily, often at inspection service request, while the service verifies effectiveness, publicizes the recall, and classifies it in Class I, Class II, or Class III terms. Around that absence, the service holds compensating tools, including administrative detention of product in commerce for up to 20 days under the meat statute at 21 U.S.C. 672, with parallel provisions for poultry and eggs, judicial seizure and condemnation pursued through the United States Attorney, retention of adulterated product inside establishments, and withholding or suspension of the federal presence on which covered operation depends. The Food and Drug Administration, by contrast, received mandatory recall authority over food, other than infant formula under its own provision, in section 206 of the FDA Food Safety Modernization Act, exercisable after an opportunity for an informal hearing where there is a reasonable probability of serious adverse health consequences. A product on the Food and Drug Administration side of the pizza line can thus be subject to a federal recall order under that authority, while a product on the inspection service side, produced under far more frequent federal presence, cannot be recalled by direct federal order under a parallel meat, poultry, or egg products power.

Taken together, those consequences explain why jurisdiction on the dinner plate deserves attention beyond its value as a collection of odd examples. Jurisdiction decides the tempo of oversight before any problem arises, the gate through which a label passes before any sale occurs, and the legal instruments available after a problem is found. A diner crossing between two items on a plate feels none of them directly, which is precisely why the plate remains such a revealing test of the system. The food presents itself as a single meal. The law behind it presents itself as an archive of separate decisions, layered by commodity, by form, by percentage, and by amendment, and still governing, product by product, which federal hand was present while the meal was made.

Continuous inspection and the model it locks in

Continuous inspection begins with a command that leaves little room for interpretation. The Federal Meat Inspection Act does not direct the Department of Agriculture to sample slaughter, to audit slaughter records after the fact, or to visit a slaughter establishment on a schedule set by risk. The statute directs examination of the animals and the carcasses themselves, by federal inspectors, as a condition of lawful operation. The Poultry Products Inspection Act carries the same structure into poultry slaughter, and the Egg Products Inspection Act extends a presence based model to egg products plants. In each case the legal unit is not the facility in the abstract. The legal unit is the individual animal presented for slaughter, the individual carcass produced from that animal, and the parts derived from that carcass. A plant may own the building, employ the workers, purchase the livestock, and arrange the buyers, and none of those arrangements confers authority to slaughter for commerce on a given morning. Authority arrives with the inspector, because the statutes make examination part of the act of slaughter itself rather than a check applied to finished product after slaughter has ended.

Antemortem inspection gives that command its first application, before any knife work begins. Livestock presented to an establishment operating under federal inspection are examined while alive, in motion and at rest, so that animals showing signs of disease or abnormal condition can be identified, segregated, and handled under inspection direction rather than passed into the slaughter flow by establishment decision alone. The point of the live examination is separation at the moment when separation remains practical. A suspect animal held back from the line protects the carcasses that follow it and creates a record grounded in direct observation rather than in paperwork reconstructed later. Antemortem examination also disciplines the receiving side of the business. An establishment that knows every consignment will be seen alive, by a federal employee with authority to condemn or to require further examination, has reason to refuse visibly unfit animals at purchase and to keep pens, alleys, and handling in a condition that withstands observation. The statute does not ask the inspector to certify a herd, a feedlot, or a supplier. The statute asks the inspector to look at the animals actually presented, on the day of slaughter, and to decide which of them may proceed.

Postmortem inspection extends the same logic to the carcass and to every part of the carcass. After slaughter and dressing, the head, the viscera, and the carcass are presented in a manner that preserves their identity with one another until examination is complete, so that a finding in one part can determine the disposition of the whole. An abnormality in the organs can lead to condemnation of tissue that appears sound in isolation, and a sound set of organs can support passage of a carcass that requires only trimming or further handling. No carcass escapes the sequence by being one among many, and no part enters commerce on the strength of a general judgment about the lot from which it came. Retention, trimming, reexamination, and condemnation all occur within the establishment and under inspection authority, in real time, while the product remains identifiable and reachable. A carcass passed for food carries the mark of inspection. A carcass condemned leaves the food supply at the point of production rather than after distribution. A system built on examination of every carcass distributes both its protection and its cost evenly across output, regardless of the history of the establishment, the condition of a particular lot, or the relative risk presented by one species or product compared with another.

What does continuous inspection require on the slaughter line?

Continuous inspection requires examination before slaughter, examination of every carcass and its parts after slaughter, and the presence of a federal inspector whenever slaughter or dressing takes place. Without that presence an establishment has no lawful authority to slaughter, however ready its workers, equipment, and livestock may be.

The prohibition that enforces the requirement is stated in operational terms. No animal may be slaughtered and dressed unless an inspector has examined it, and inspectors are on the line during all operating hours of a slaughter establishment. Presence is therefore not measured by a visit, a tour, or a daily certification at the end of a shift. Presence means a federal inspector positioned to observe, to retain product for further examination, to require corrective action, and to stop questioned product from moving forward as inspected and passed. An establishment cannot lawfully begin early, run late, or add a shift on its own authority and settle the inspection question afterward. The line runs when coverage runs, and the line does not lawfully run when coverage is absent. That single linkage explains much of the administrative character of the meat and poultry system. Scheduling becomes a matter of federal staffing as well as plant management. A vacancy, an illness, or a gap in relief coverage is not merely a personnel problem inside an agency office. The same gap can become lost production inside a private establishment, with livestock already delivered and workers already assembled, against output that cannot legally be produced until coverage is restored.

The dependence runs in the opposite direction as well. The inspection service must organize a workforce around the operating calendars of establishments whose hours follow livestock deliveries, perishability, labor agreements, and market demand rather than a uniform government schedule. Inspectors rotate, relieve one another, and cover absences so that the promise of presence during all operating hours holds in practice. An establishment, for its part, plans line speeds, staffing, and maintenance around the coverage assigned to it. Over time the arrangement hardens into expectation on both sides. The agency staffs to the plant calendar that exists, and the plant builds capacity, cooler space, labor model, and customer commitments around the coverage the agency provides. Both sides experience the arrangement as shared infrastructure, closer to a utility that must arrive every morning than to an audit that arrives on a cycle.

Processing establishments occupy a different position inside the same statutory family, and the difference clarifies what continuous has meant in practice. A plant that grinds, cures, cooks, cans, or otherwise processes inspected meat or poultry does not present a live animal or a newly dressed carcass for the sequence of antemortem and postmortem decisions that defines slaughter. The standard model for such an establishment provides inspection presence at least once per operating day. The Congressional Research Service has described processing plants visited once every day as treated as under continuous inspection in keeping with the inspection laws, a formulation that captures both the breadth of the statutory claim and the practical adjustment inside it. Daily presence keeps a federal inspector in regular contact with sanitation, formulation, labeling, handling, and records while product moves through the plant, and daily presence distinguishes the meat and poultry regime from a regime of occasional facility visits.

The methods inspectors brought to the line for most of the twentieth century grew from the world the original statutes confronted. The inheritance was organoleptic, grounded in sight, smell, and touch. An inspector looked for discoloration, lesions, swelling, parasites, abnormal tissue, and visible contamination. An inspector smelled for odors that signaled decomposition or disease. An inspector palpated tissue, incised glands and organs where examination required it, and judged consistency by hand. Those senses, trained by veterinary education and by repetition across thousands of carcasses, were well matched to the hazards that dominated the early meat supply. Condemned animals and carcasses in that era often presented defects that a trained observer could see, feel, or smell, including tuberculosis lesions, abscesses, emaciation, and filth carried into the plant or created during dressing. A system designed around visible disease and visible insanitation naturally placed its trust in the observer stationed where the defects appeared. Lines moved at speeds that allowed handling and incision, establishments were smaller, and the distance between a federal decision and the individual carcass was short enough for the senses to serve as the primary instrument.

The limits of the inheritance follow from the same history. Microbial contamination that produces severe illness often leaves no lesion, no odor, and no change in texture that sight, smell, or touch can detect. A carcass can appear clean, smell normal, and feel sound while carrying organisms capable of causing disease, and ground product can combine material from many carcasses in a manner that spreads an invisible problem across a large volume of food. Organoleptic examination retains value for the defects it was built to find, including visible contamination that signals a breakdown in dressing or sanitation. The method nevertheless answers a different question from the question posed by invisible pathogens. Examination of appearance asks whether a carcass shows defect. Microbiological safety asks whether a process, tested by sampling and controlled by preventive systems, keeps contamination within bounds that appearance cannot measure. The tension shaped the rulemaking struggles of the inspection service for decades, because the statute continued to command presence and carcass examination even as the case for supplementing the senses with laboratory testing and process control grew stronger. Presence could be redirected and supplemented. Presence could not simply be withdrawn, because the statutory command for examination of every carcass remained in force.

Staffing gives the model its measurable scale. A Congressional Research Service issue brief from the period around 2003, Issue Brief IB10082, counted about 7,700 Food Safety and Inspection Service field employees, roughly 1,000 of them veterinarians, serving some 6,200 plants and import stations. Those figures belong to that source and that period, and the value of the figures lies in the structure they reveal rather than in any claim about a later headcount. A presence based system requires a field force large enough to station personnel in establishments across the country, across shifts, and across the geography of livestock production, which concentrates slaughter in some regions and disperses processing across many more. Veterinarians form a comparatively small share of the force and carry responsibilities that include disease disposition, humane handling oversight, and supervision of inspectors whose daily work is examination and enforcement. The ratio embedded in the figures describes a pyramid. A broad base of inspectors performs the continuous work the statutes command, a narrower layer of veterinary and supervisory personnel supports judgment calls that exceed routine disposition, and the entire structure exists because the law ties lawful production to federal attendance. Retain the attendance requirement and workforce planning becomes production planning for a substantial share of the national meat supply.

The cost structure follows the same linkage, divided between public appropriation and establishment payment according to the character of the hours worked. Inspection during an establishment normal operating schedule is a public function funded largely by congressional appropriations, reflecting the statutory judgment that inspected product is the condition of lawful commerce rather than a private service purchased at will. Inspection beyond normal hours operates on a different footing. An establishment that seeks coverage for overtime, for work on days outside the regular schedule, or for hours added to an assigned tour requests a service made necessary by its own production choices, and the fee structure for such coverage places the added cost on the establishment that generates it, under procedures set by regulation rather than by negotiation between a plant and an individual inspector. Voluntary inspection for species and products outside the mandatory acts follows the fee for service pattern more completely, with the recipient paying for a service the law does not command. The distinctions matter for incentives. Appropriated coverage treats the baseline as a public guarantee attached to the mark. Reimbursable overtime treats expansion beyond the baseline as a business expense, visible in plant accounting and capable of influencing decisions about shifts, weekends, and seasonal peaks.

No comparison exposes the distinctiveness of the meat and poultry model more sharply than the model Congress gave the Food and Drug Administration for the balance of the food supply. The Government Accountability Office, in GAO-17-74 from 2016, described the Food and Drug Administration as responsible for roughly four fifths of the United States food supply and the Food Safety and Inspection Service as responsible for roughly one fifth, while the two agencies received close to the same food safety funding. The funding comparison reflects the labor intensity of presence. The Food and Drug Administration regulates food facilities through registration, standards, and periodic inspection directed by risk, rather than through attendance during operating hours. Under the FDA Food Safety Modernization Act, enacted in 2011, high risk domestic facilities are inspected at least once every three years and other domestic facilities at least once every five years, frequencies stated in the statute itself. A facility under that regime operates on most days without any federal inspector inside the building, maintains preventive controls under its own management, and answers for compliance through records, sampling, and inspections that arrive at intervals measured in years. A slaughter establishment under the meat statute operates under the opposite assumption. Federal attendance is the normal condition of production, intervals are measured in shifts rather than years, and the absence of an inspector suspends lawful operation rather than merely postponing an oversight encounter.

What continuous presence buys follows directly from its position inside production. The first purchase is real time condemnation. Product found diseased, contaminated, or otherwise unfit can be retained, trimmed, reexamined, or condemned at the moment of finding, before commingling, packaging, storage, or shipment makes separation costly or impossible. The second purchase is deterrence through visibility. Workers, supervisors, and managers perform dressing, sanitation, and handling under observation by an official who can act on the spot, and the knowledge that observation is constant rather than occasional shapes conduct in ways that announced visits cannot reproduce. The third purchase is the mark of inspection itself, which functions as a passport in commerce. The legend applied to passed product signals to distributors, retailers, institutional buyers, export customers, and consumers that each carcass behind the product stood within a federal examination sequence, and the signal travels with the product into markets where the buyer has no independent means of judging the conditions of slaughter. A buyer in one state can accept product slaughtered in another state without inspecting the plant, because the federal mark carries the assurance that a common statutory process was applied. The passport value also explains resistance to any proposal perceived as weakening presence. A change framed by an agency as modernization can be heard by buyers, workers, and members of Congress as a change to the meaning of the mark itself.

The costs of the model are the mirror image of the purchases. Rigidity stands first. A statute that commands examination of every carcass and attendance during all operating hours leaves an agency limited room to shift resources toward the establishments, products, or processes presenting the greatest risk at a given time. A plant with a strong compliance history and a plant with repeated problems both require coverage for the line to run, and the baseline demand of the strong plant does not fall simply because supervision might yield more protection elsewhere. Resources are spent regardless of relative risk because the entitlement to coverage and the obligation to provide coverage attach to operation itself. The Government Accountability Office criticized exactly that allocation pattern across a series of reports on federal food safety oversight, faulting a structure in which inspection resources follow statutory mandate rather than risk. The criticism asks whether the marginal inspector hour produces the same protection in every setting, and the continuous model, by design, has difficulty answering in comparative terms. Hours are allocated to keep lines lawful. Risk based allocation would allocate hours to reduce expected illness. Those are different objectives, and a workforce sized for the first objective cannot be redeployed to the second without statutory change and a visible loss of presence in establishments that built operations around guaranteed coverage.

A second cost lies in the pull the model exerts on every reform placed on top of it. Preventive controls, sampling programs, sanitation procedures, and performance standards must all be fitted around a carcass by carcass examination sequence that predates them, staffed by a workforce whose positions, tours, and duties were defined by that sequence. Reform therefore tends to arrive as addition rather than substitution. New duties accumulate for inspectors who remain responsible for the original examination, and establishments carry both the older presence based requirements and the newer process based requirements at once. The optional modernization systems illustrate the boundary of permissible change within the statutes as they stood before February 2021. The Food Safety and Inspection Service finalized a new swine slaughter inspection system in 2019 and a poultry slaughter modernization rule in 2014, each optional, each shifting some preliminary sorting tasks to establishment employees, and each preserving the inspector role in carcass by carcass determination. Rulemaking can rearrange who sorts, who examines first, and what additional controls apply. Rulemaking under the existing statutes cannot convert slaughter inspection into the periodic, risk based facility regime that governs most other food, because the conversion would contradict the command that structures the entire program.

Imports test the model at the border, where continuous presence inside a domestic establishment cannot be reproduced in the same form. The Food Safety and Inspection Service addresses foreign production through equivalence. A country seeking to export meat, poultry, or egg products to the United States must maintain an inspection system equivalent to the United States system, judged by the inspection service through document review, on site audit, and continuing evaluation, and only establishments certified within an eligible country system may ship. Equivalence extends the logic of the domestic model outward. The exporting country supplies the continuous presence under its own inspectors, and the United States accepts the resulting product because the foreign system is held to deliver protections equivalent to the protections delivered by federal attendance in a domestic plant. At entry, imported product is subject to reinspection at import establishments, where inspectors examine shipments, verify certification and labeling, and select product for further examination or sampling before release into commerce. Product refused entry does not receive the mark and does not proceed as inspected product. The double structure, equivalence abroad and reinspection at arrival, compensates for the absence of a United States inspector on a foreign line.

The model locked in by continuous inspection is therefore larger than a technique for examining carcasses. The technique produced an institution, a workforce, a funding pattern, a commercial passport, and a set of expectations that later lawmakers and administrators inherited as fixed conditions. Support for the model, stated at strength by consumer organizations, inspection workforce representatives, and members of Congress who defended the mark, rests on immediacy. Hazards found at slaughter are stopped at slaughter, conduct is observed rather than inferred, and the assurance behind the mark rests on examination of the product itself. Skepticism about the model, stated at strength by the Government Accountability Office and by administrations seeking flexibility across several decades, rests on allocation. A mandate that spends inspector time wherever a line operates spends less time where risk concentrates, and a fragmented federal food safety structure makes the rigidity of the largest presence based program a central obstacle to a unified risk based strategy. Neither position displaces the statutory text. The text commands antemortem examination, postmortem examination of every carcass and its parts, and attendance during operating hours, and every administration through February 2021 administered that command with a workforce organized to fulfill it.

The outbreak that rewrote the rulebook

In January 1993, an outbreak of Escherichia coli O157:H7 infection was traced to undercooked ground beef hamburgers served by the Jack in the Box restaurant chain. The patties had been supplied by a California processor, and the illnesses that followed were not first recognized as a single foodborne event scattered across restaurant meals. Detection came in Washington State, through a cluster of children with hemolytic uremic syndrome, a severe complication that turned individual medical cases into a pattern investigators could follow. That route of detection mattered to everything that came afterward, because the signal reached public health authorities through gravely ill children rather than through routine product testing or establishment records.

The toll fixed the scale of the failure in public and institutional memory. The Centers for Disease Control and Prevention, in its retrospective account, put illnesses at more than 700 people across four states, Washington, Idaho, California, and Nevada, and recorded that four children were killed. The geographic spread followed the distribution of a processed product through a restaurant system, so a hazard formed at the level of raw material and preparation produced consequences in state after state. A localized cooking error at one outlet could not explain the pattern. The product itself, and the system that had passed it into commerce, became the objects of inquiry.

The cluster also exposed a mismatch between the hazard and the tools built to find hazards in meat plants. Hemolytic uremic syndrome in children pointed investigators toward a toxin producing organism in ground beef, yet nothing about that organism announced itself to sight, smell, or touch at the point of production. Ground beef could carry Escherichia coli O157:H7 and present no visible spoilage, no off odor, and no texture change that an inspector on a slaughter line or in a grinding room could detect by handling the product. The outbreak therefore posed a question that went beyond one processor or one restaurant chain, reaching the premises of the federal meat program itself.

That program had been built on organoleptic examination. Under the Federal Meat Inspection Act, inspectors examined livestock before slaughter and examined every carcass and its parts after slaughter, and slaughter establishments could not lawfully operate without inspection coverage during operating hours. Presence was the guarantee the statute supplied. An inspector was on the line, carcasses were looked at and handled, and diseased or visibly unwholesome material could be identified and kept from commerce. The model addressed conditions that examination could reveal, and it placed federal personnel inside the establishment as a condition of operation.

Against a bacterial contaminant, that presence offered limited protection. A carcass could pass postmortem examination and still carry pathogenic bacteria in numbers no examination could register, and grinding then distributed surface contamination through a batch of product rather than leaving it at a single point. The Jack in the Box outbreak made that limitation concrete in the most damaging form available, through hamburgers eaten by children and through a complication that physicians recognized before the food source was identified. Confidence in sensory examination as a sufficient answer to meat safety sustained damage that ordinary program criticism had not inflicted, because the failure was measured in deaths rather than in audit findings.

The institutional shock worked through law as much as through public attention. If examination could not see the hazard, the Food Safety and Inspection Service needed a legal category that made the hazard unlawful in the product whether or not an inspector had observed a defect, and a control method that operated on process rather than on appearance. The adulterant concept in the meat inspection act supplied the first element. A preventive control system, imposed on establishments and verified by the agency, supplied the second. The sequence from outbreak to new rulebook ran through both elements, in that order, over the seven years that followed detection in Washington.

How did a restaurant outbreak change the legal status of a pathogen in ground beef?

The January outbreak led the inspection service to treat Escherichia coli O157:H7 in raw ground beef as an adulterant under the meat inspection act in September of the following year. That legal reclassification made contaminated product unlawful to sell and led, through litigation and rulemaking, toward mandatory hazard control systems in slaughter and grinding establishments.

The declaration came on September 29, 1994, from Michael R. Taylor, Administrator of the Food Safety and Inspection Service, serving under Secretary of Agriculture Mike Espy. Taylor announced that the service would treat Escherichia coli O157:H7 in raw ground beef as an adulterant under the Federal Meat Inspection Act. The attribution matters to the institutional record. The policy was an administrator level exercise of the inspection statute, issued through the agency charged with administering that statute, and not a personal pronouncement detached from the rulemaking and enforcement machinery that would have to sustain it in court and in establishments across the country.

Adulterant status changed the legal consequence of finding the organism. Before the announcement, contamination of raw ground beef with a pathogen could be treated as a problem that cooking was expected to resolve, a framing that left much of the protective burden at the restaurant grill and the household stove. After the announcement, raw ground beef bearing Escherichia coli O157:H7 was product that the statute did not permit to be sold. The finding itself carried the consequence, without any need to show that a particular inspector had seen filth, disease, or mishandling in the establishment that produced a particular batch. Sampling results could therefore do work that sensory examination could not do.

Industry resistance followed promptly, and it took legal form. Industry challenged the policy in Texas Food Industry Association v. Espy, contesting the authority of the inspection service to declare the organism an adulterant in raw ground beef. The federal district court refused to block the policy in 1994, and the Fifth Circuit affirmed that result in 1995. The litigation therefore settled, at the level of a governing appellate decision in the circuit where the challenge was brought, that the administrator announcement was not an empty statement of intent. The policy survived the forum industry had chosen for testing it, and the sampling and rulemaking steps that followed rested on that survival.

The reasoning that sustained the policy turned on how ground beef is actually eaten. Per the reasoning in the Espy litigation, many consumers regard ground beef cooked rare or medium as properly cooked, so ordinary cooking practice does not reliably destroy the organism. That factual premise did decisive legal work. A pathogen that proper cooking reliably kills occupies one position under a statute built around consumer handling of raw meat. A pathogen that survives preparation methods consumers treat as finished cooking occupies another, because the assumption that the kitchen completes the safety system fails on the evidence of ordinary behavior at the stove and the grill.

The distinction also separated Escherichia coli O157:H7 from pathogens that the legal system had long treated as incidents of raw product. Where an organism is destroyed by the cooking that consumers understand raw meat to require, presence in the raw product and injury after proper preparation do not align. In the ground beef setting, the alignment failed in the opposite direction. Product served in the manner many consumers considered proper could still deliver a live organism with a low infectious dose, and the outbreak record from January 1993 showed where that delivery could end. The courts accepted a definition of adulteration grounded in consumption as practiced, not in consumption as instructions imagined it.

While the challenge moved through the courts, the inspection service built the measurement system the new status required. In October 1994, following the Taylor announcement, the Food Safety and Inspection Service began sampling raw ground beef for Escherichia coli O157:H7. Sampling converted the declaration from a legal proposition into an operating program. Product could be tested, positives could be identified as adulterated, and establishments grinding beef faced a defined federal check directed at the specific organism rather than at general appearance or general sanitation alone. The order of events preserved the logic of the chain. Status came first, testing gave status practical effect, and litigation confirmed that the sequence would not be unwound at its first step.

A sampling program aimed at one organism in one product, however, addressed only the narrowest part of the problem the outbreak had revealed. Ground beef was the vehicle in 1993, but the underlying defect was a production system that had no general obligation to identify microbial hazards in advance, control the points where contamination could be prevented or reduced, and document that those controls had operated. A test for a single pathogen could find failures after production. It could not, by itself, require an establishment to organize production so that failures became less likely in the first instance. That gap between detection and prevention supplied the reason for the rulemaking that followed.

In February 1995, the Food Safety and Inspection Service proposed the Pathogen Reduction and Hazard Analysis and Critical Control Point rule, known in shortened form as the Pathogen Reduction and HACCP rule. The proposal generalized the response. Instead of resting federal policy on adulterant findings in finished ground beef alone, the proposed rule would require establishments to build sanitation and process controls into daily operation and to adopt hazard analysis systems as a condition of inspected production. The proposal thus linked two ideas that the outbreak had forced together, pathogen reduction as a measurable objective and hazard control as the establishment level method for reaching that objective before product left the plant.

The final rule issued on July 25, 1996, as Pathogen Reduction; Hazard Analysis and Critical Control Point (HACCP) Systems, published at 61 FR 38806. Its structure had four components, and the components worked as a set rather than as interchangeable parts. Sanitation Standard Operating Procedures supplied the baseline conditions of plant hygiene. Generic Escherichia coli testing supplied a process control indicator in slaughter establishments. Salmonella performance standards supplied pathogen reduction benchmarks for slaughter establishments and for raw ground product. Mandatory HACCP systems supplied the establishment specific plan that tied hazard identification to control points in a particular operation. Removing any one component changed the character of the other three.

Sanitation Standard Operating Procedures, or Sanitation SOPs, addressed the environment in which production occurred. An establishment had to state, in operating procedure form, how sanitation would be carried out, rather than treating cleanliness as an informal habit that inspectors assessed only by walking the floor. The mechanism was unglamorous and foundational. Contamination pathways run through equipment, surfaces, employees, and plant routines, and a written sanitation procedure gave the inspection service a stated standard against which daily practice could be checked. Failure could then be located in a defined procedure and corrected as a system fault, instead of being recorded as a general impression that conditions in a room appeared unsatisfactory on a given morning.

Generic Escherichia coli testing addressed process control rather than the specific outbreak organism. The rule required testing for generic Escherichia coli in slaughter establishments as an indicator of whether slaughter and dressing processes were controlling fecal contamination. The distinction between the generic organism and the O157:H7 strain carried the design of the program. The adulterant policy pursued a named pathogen in raw ground beef. The process control test measured the performance of the slaughter process itself, using an organism associated with fecal contamination as the signal. One element policed a product. The other element told the agency and the establishment whether the process producing carcasses was holding within expected bounds.

Salmonella performance standards supplied the third element, a benchmark expressed as prevalence in tested product. The final rule set pathogen reduction performance standards for Salmonella in slaughter establishments and in raw ground product, so that an establishment could be measured against an expected level of reduction rather than judged only on whether any single sample returned a positive result. In the architecture of the 1996 rule, the standards functioned as verification measures within a broader control system. Performance against the benchmark told the inspection service how an establishment process was performing over a series of samples, and persistent failure signaled that controls required attention through the sanitation and hazard systems the rule had made mandatory.

Mandatory HACCP systems formed the fourth and most consequential component. HACCP, built from hazard analysis and critical control points, required an establishment to analyze its own operation for hazards reasonably likely to occur, identify the points in production where those hazards could be controlled, and operate controls at those points. The requirement displaced a model in which federal personnel effectively defined safety through continuous observation of product and plant. Under the rule, the plan belonged to the operation that produced the food, because only that operation controlled the equipment, the suppliers, the line speeds, the interventions, and the corrective decisions that determined whether a hazard was managed on a particular day.

The phase in calendar translated the size of that change into dates that varied with establishment capacity. Sanitation SOPs and generic Escherichia coli process control testing took effect on January 27, 1997, placing the hygiene baseline and the slaughter process indicator in force before the full hazard system obligation arrived. The staging reflected a practical sequence. An establishment first documented sanitation and entered the testing regime, then assumed the more demanding duty of designing and running a complete hazard plan. The rule therefore entered production in layers, and each layer gave the inspection service records and results to examine before the next layer imposed its own set of verification tasks.

HACCP itself became mandatory on January 26, 1998, for large establishments, defined in the rule by a threshold of 500 or more employees. Small establishments, with 10 to 499 employees, came under mandatory HACCP on January 25, 1999. Very small establishments, with fewer than 10 employees or under 2.5 million dollars in annual sales, came under the requirement on January 25, 2000. The staggered calendar distributed the compliance burden across three Januarys and acknowledged a difference in administrative capacity between a plant with dedicated technical staff and a small operation where plan design competed with every other duty of ownership. By the final date in the sequence, the hazard system obligation covered the inspected sector as a whole rather than only its largest firms.

The institutional change inside those dates deserves attention apart from the calendar. Before the rule, the federal role in a slaughter establishment centered on performance by government personnel, the antemortem and postmortem examinations and the daily presence through which inspectors observed conditions directly. After the rule, the establishment owned a written hazard plan and bore responsibility for operating it. The agency role shifted toward verification. Inspection personnel examined whether the plan addressed the hazards of the operation, whether controls at identified points were being carried out, whether records showed the system functioning, and whether deviations drew corrective responses from the establishment itself.

Verification is a different discipline from performance, and the difference redistributed expertise as well as labor. An establishment could no longer treat safety as a condition conferred by the presence of an inspector at the rail, because the rule assigned the analytical task to the producer. Hazard identification, control selection, monitoring, and correction became management functions inside the plant, subject to federal check. At the same time, the inspection service had to evaluate systems rather than only carcasses, reading plans and records and testing product as evidence of how a system performed. The outbreak chain thus produced a durable reallocation. The producer generated the safety system. The government tested whether that system deserved confidence.

That reallocation also changed how failure would be described in later disputes. Under a purely sensory model, failure appeared as a condition an inspector had missed. Under a hazard system model, failure could be located in the design of a plan that omitted a hazard, in monitoring that did not detect a deviation, or in a corrective response that allowed affected product to proceed. Accountability followed the plan. Because the establishment wrote the plan and ran the controls, gaps in either document or execution belonged to the establishment in the first instance, while the agency answerability ran to the rigor of verification. The 1996 rule therefore altered not only production practice but the language in which responsibility for a later contamination event would be assigned.

The chain did not close in January 2000, when the last size category entered mandatory HACCP. The adulterant policy that Taylor had announced for raw ground beef in 1994 reached a broader category of product in 1999, when the Food Safety and Inspection Service extended the policy to non-intact beef products. The extension followed the same logic as the original declaration. Beef whose structure had been disrupted presented a different risk position from an intact cut, because contamination associated with the surface could no longer be treated as remaining only where ordinary preparation of a whole cut would address it. Extending adulterant treatment to the non-intact category aligned the legal category with the physical character of the product.

A second later link tested the hazard plans that establishments had written under the 1996 rule. In 2002, after a large ground beef recall, the inspection service ordered ground beef establishments to reassess their HACCP plans for Escherichia coli O157:H7. Reassessment used the ownership structure the rule had created. Because each establishment held its own plan, the agency could direct every plan holder to revisit a specific hazard in light of operating experience and recall evidence, rather than rewriting a single government procedure for the entire sector. The order treated a hazard analysis as a living part of the system, capable of becoming inadequate when experience showed that a hazard had been underestimated or a control had not performed as the plan assumed.

The reassessment order also showed the limit built into the verification model. A written plan could satisfy formal requirements and still misjudge the hazard that mattered most in a particular grinding operation, and verification by the agency depended on pressing establishments to confront that possibility directly. The 2002 step therefore operated as a correction inside the system rather than as a replacement of it. The service did not return to defining safety through sensory examination, and it did not abandon establishment ownership of plans. It used the authority the HACCP structure supplied to require a fresh examination of the hazard at the center of the 1993 outbreak, nine years after detection and six years after the final rule had issued.

The last link in the sequence widened the list of organisms treated as adulterants beyond the single strain named in 1994. In September 2011, the Food Safety and Inspection Service announced that six additional Shiga toxin producing Escherichia coli strains, O26, O45, O103, O111, O121, and O145, would be treated as adulterants in raw non-intact beef. The category of product matched the 1999 extension, while the category of organism expanded from Escherichia coli O157:H7 alone to a group of related strains sharing the Shiga toxin producing character that had made the original strain a concern for severe illness. Policy again moved by reclassification, assigning a legal consequence to laboratory findings in a defined product class.

Routine testing for the six additional strains began on June 4, 2012. The interval between the September 2011 announcement and the June 2012 start of testing repeated, at smaller scale, the pattern set in 1994 and 1995, when a declaration preceded the sampling program that gave it routine effect. Announcement established the legal position. Testing operationalized that position in laboratories and establishments. By February 2021, the sequence that began with children hospitalized in Washington in January 1993 stood as a connected set of mechanisms rather than as a single reform moment, each step supplying the premise for the step that followed it.

Taken end to end, the chain ran from detection through reclassification to system building, and then back through reclassification as knowledge and operating experience accumulated. A syndrome cluster identified the vehicle. An administrator declaration made the vehicle pathogen an adulterant. A court decision sustained that move on the ground of actual consumer cooking practice. A sampling program made the status testable. A proposed and then final rule converted a product specific response into establishment wide duties for sanitation, process control testing, pathogen performance, and hazard planning. Later extensions widened the product category, a reassessment order refreshed the plans, and a second adulterant announcement widened the organism list for non-intact beef.

Nothing in that sequence restored the premise that had failed in 1993. Federal inspection after the Pathogen Reduction and HACCP rule retained carcass examination and plant presence under the meat inspection act, but safety for ground beef and non-intact beef no longer rested on examination alone. It rested on a legal status for specified organisms, on testing directed at those organisms and at process indicators, and on establishment plans that the agency verified rather than performed. The rulebook that emerged from the outbreak was therefore not a thicker version of the sensory manual. It was a different allocation of duties, written in response to a hazard that examination could not see and that ordinary cooking, as consumers practiced it, did not reliably eliminate.

The pathogen the agency cannot reach: Salmonella and the limits of the statute

The scale of Salmonella illness frames every legal question that follows. The Centers for Disease Control and Prevention estimates about 1.35 million Salmonella infections, 26,500 hospitalizations, and 420 deaths in the United States each year. Those figures form the agency’s standing estimate, and they place a familiar organism in an unfamiliar position for a food safety regulator. The pathogen causing the largest measured burden in the meat and poultry supply was the pathogen the inspection service had the least direct power to act against. The obstacle sat in the statutory definition of adulteration and in a judicial reading of that language that separated the presence of an organism from the legal consequences that presence could carry.

The burden estimate rests on a specific study as well as on the standing figure. Scallan and coauthors, in a study published in Emerging Infectious Diseases in 2011, estimated about 1.03 million domestically acquired foodborne nontyphoidal Salmonella illnesses each year in the United States. The two figures measure related but distinct things. The standing estimate counts Salmonella infections broadly, while the Scallan estimate isolates illnesses acquired domestically through food. Read together, the figures show that the foodborne share alone exceeded one million illnesses a year. In the Centers for Disease Control and Prevention burden estimates, Salmonella is the leading bacterial cause of foodborne hospitalization and death.

Frequency changed the meaning of the legal limit. Salmonella forced a reckoning because prevalence in raw product and illness in the population coexisted with a settled legal position that the organism, by itself, did not make raw meat or poultry adulterated. The Food Safety and Inspection Service therefore confronted a gap between measurement and authority. Measurement improved through the nineteen nineties, as sampling programs generated establishment level data on Salmonella prevalence by product class. Authority did not move with measurement. The data could show that one establishment’s product carried the organism far more often than the product of comparable establishments, yet the data alone could not supply the statutory predicate for the strongest enforcement action the service possessed, the withdrawal of inspection that closes an establishment.

The old settlement came first, and the later conflict grew out of it. In American Public Health Association v. Butz, 511 F.2d 331, decided by the Court of Appeals for the District of Columbia Circuit in 1974, the court held that Salmonella in meat does not itself render the product adulterated. The reasoning turned on two connected propositions about raw meat and the people who buy it. Ordinary cooking destroys the organism. Buyers know that raw meat requires cooking, and raw meat reaches the kitchen as an ingredient to be prepared rather than as a food to be eaten as purchased. On those propositions, the presence of Salmonella in the raw product did not make the product unfit or injurious in the legal sense the plaintiffs pressed, because the ordinary and expected step between purchase and consumption removed the hazard. The court treated the organism as an inherent feature of raw meat that customary preparation addressed, rather than as a defect that the inspection statute condemned in the product as sold.

The Department of Agriculture shared that position for decades, and the agreement matters as much as the holding. The department’s own stance, maintained across the period after the District of Columbia Circuit decision, was that Salmonella is not an adulterant per se. A per se adulterant occupies a distinct place in meat inspection law. Once the service declares an organism or substance an adulterant in a defined product, product bearing that organism or substance is adulterated as a matter of law, subject to retention inside the establishment, to refusal of the marks of inspection, and to the enforcement sequence that follows from an adulteration finding. Salmonella never received that designation for raw meat and poultry. Detection of the organism therefore produced information, and information could prompt further inquiry, closer scrutiny, and demands for corrective action grounded in some other requirement. Detection alone produced no adulteration finding, and without an adulteration finding the most direct route from a laboratory result to a legal violation remained closed.

The settlement also rested on an assumption about where control belonged. If cooking at the point of consumption destroyed the organism, then the burden of the final kill step sat in homes and restaurants, while the establishment’s obligation centered on sanitation and on preventing avoidable contamination. The settlement accommodated prevalence. It treated the organism as a condition of the raw material to be managed downward and neutralized at cooking, rather than as a contaminant whose any presence proved a plant failure.

The Pathogen Reduction and Hazard Analysis and Critical Control Point rule, issued as a final rule on July 25, 1996, at 61 Federal Register 38806, changed the measurement side of the settlement without changing the underlying definition of adulteration. The rule had four components: Sanitation Standard Operating Procedures, generic Escherichia coli process control testing in slaughter establishments, Salmonella pathogen reduction performance standards for slaughter establishments and raw ground product, and mandatory Hazard Analysis and Critical Control Point systems. The Salmonella standards sat inside a larger redesign. Sanitation procedures addressed the daily conditions of the plant. Process control testing for generic Escherichia coli gave slaughter establishments an indicator organism for fecal contamination control. Hazard analysis systems required each establishment to identify hazards, set critical limits, monitor them, and correct deviations. The performance standards added an outcome measure over that structure, a way for the service to test whether the combined system in a given establishment was in fact holding Salmonella prevalence down.

Design explains both the strength and the vulnerability of the standards. A performance standard of this kind operates as a proxy for process control. The standard does not declare each positive sample an adulterated product, and the rule did not purport to make Salmonella an adulterant. Instead, the standard set a level of prevalence, measured across a set of samples drawn from an establishment’s production, that the establishment was expected to meet for its product class. An establishment whose samples exceeded the permitted proportion of positives had, by the logic of the design, failed to control its process as comparable establishments could. Prevalence varied by product class, because raw material, animal source, and processing steps differed, so the standards were graded by product class rather than imposed as one uniform figure across all raw product. Ground product, carcasses, and other classes each carried a standard fitted to the baseline prevalence and the achievable reduction for that class.

Graded standards also allowed the service to press for reduction without asserting that product above the standard was unsafe to eat after cooking. Establishments accepted outcome testing and the obligation to meet the prevalence level for the class, and the service gained a uniform, repeatable measure that exposed establishments whose sanitation and process controls lagged. Failure showed a process out of control. What legal consequence failure could carry remained a separate question, and the rule’s enforcement provisions supplied the answer the service intended to give: repeated failure could lead the service to withhold the marks of inspection and to move toward withdrawal of inspection. That intended consequence became the point of litigation when an establishment that had failed repeatedly refused to accept closure on that basis.

Supreme Beef Processors supplied the test case. Supreme Beef Processors, Inc. was a ground beef processor in Dallas, Texas, and it supplied ground beef to the National School Lunch Program. The company therefore operated under federal inspection and also sold into a federal feeding program with its own purchasing interest in the safety of the product. Under the performance standard for ground beef, the service drew sample sets from the establishment’s production and tested for Salmonella. Supreme Beef failed the standard repeatedly. In one round of testing, about 47 percent of samples were positive for Salmonella, a proportion far above the level the standard permitted for the product class. The sequence of failures, rather than a single aberrant set, formed the factual predicate for the enforcement step that followed, because the design of the standard treated repeated failure across sample sets as the signal that process control had not been restored after notice of earlier results.

The enforcement step was the most serious one available to the service short of an adulteration action against specific product. The Food Safety and Inspection Service moved to withdraw inspection from the establishment. Withdrawal of inspection is not a fine and not a warning. A meat establishment under the Federal Meat Inspection Act cannot lawfully operate without inspection, and product cannot bear the marks of inspection or enter commerce as inspected product without the service’s presence and approval. Withdrawal therefore closes the plant as effectively as a padlock, even though its legal form is the removal of a government service the statute requires before operation. Supreme Beef sued the Department of Agriculture to block the withdrawal. The district court held for the company, and the department appealed to the Court of Appeals for the Fifth Circuit, the circuit that includes Texas. The question presented, stripped of its surrounding disputes, asked whether failure of the Salmonella performance standard, standing alone, gave the service statutory authority to withhold the inspection marks and withdraw inspection from the establishment.

Why could standard failure alone not support withdrawal of inspection?

The Fifth Circuit held that Salmonella is neither an added substance that renders product injurious nor the product of insanitary conditions that rendered product injurious, because normal cooking destroys the organism and the statute’s word rendered requires a deleterious change during preparation, packing, or holding inside the establishment.

The full reasoning in Supreme Beef Processors, Inc. v. United States Department of Agriculture, 275 F.3d 432, decided by the Fifth Circuit in December 2001 with Judge Higginbotham writing for the court, proceeded in two branches, one for each statutory route the service might travel to reach adulteration. The first branch ran through 21 U.S.C. 601(m)(1). Under that provision, a product is adulterated if it bears or contains an added substance that may render the product injurious to health, among the other grounds the paragraph states. The court addressed the elements in order. Salmonella, the court reasoned, is not an added substance in the relevant sense. The organism occurs naturally in a substantial proportion of raw meat and poultry, arriving with the raw material itself rather than entering product through some addition during processing. A substance inherent in the raw material at meaningful prevalence across the industry did not fit the statutory category of a substance added to the product.

The second element of the first branch failed on independent ground. Even a substance within the paragraph must be one that may render the product injurious to health, and the court reasoned that Salmonella’s presence does not render raw ground beef injurious to health because normal cooking practices destroy the organism. The reasoning tracked the District of Columbia Circuit’s reasoning in the Butz case and extended it into the performance standard context. Raw ground beef is sold for cooking, consumers cook it, and cooking at ordinary temperatures kills Salmonella. A hazard that the ordinary and expected preparation of the product eliminates does not leave the product injurious in the statutory sense. The combination of the two points closed the first branch. The organism was not added, and its presence did not render product injurious, so product bearing Salmonella at the levels the samples showed was not adulterated under 601(m)(1), and a standard measuring the prevalence of a nonadulterant organism could not create adulteration where the paragraph did not.

The second branch ran through 21 U.S.C. 601(m)(4). Under that provision, a product is adulterated if it has been prepared, packed, or held under insanitary conditions whereby it may have become contaminated with filth, or whereby it may have been rendered injurious to health. The service’s theory treated excessive Salmonella prevalence as evidence that the establishment’s conditions were insanitary and that product had accordingly been rendered injurious. The court read the verb rendered as the controlling limit. Rendered, in the court’s reading, requires a deleterious change in the product, a change brought about during preparation, packing, or holding inside the establishment. Product must have been made injurious by something that happened there. The Salmonella content of incoming raw material is not a change worked inside the plant. The organism arrives with the animals and the meat, at levels set by conditions upstream of the establishment, and its prevalence in incoming material is a characteristic of that material. A characteristic the product brings through the door, the court reasoned, is not a condition the statute lets the service regulate through the performance standard under this paragraph, because the paragraph reaches product rendered injurious in the establishment, not product whose inherent microbial load the establishment failed to reduce to a prescribed level.

Cross-contamination could not repair the second branch, and the court’s treatment of that argument completes the reasoning. Within any processing establishment, product contacts surfaces, equipment, and other product, and organisms move among them. The service could point to that movement as an insanitary condition arising inside the plant. The court answered that cross-contamination cannot itself supply the insanitary condition that renders product injurious when the organism spread by that contact is not injurious after cooking. The chain in the paragraph had to end in product rendered injurious to health, and an organism that normal cooking destroys does not reach that end point, whether the organism sat on the product on arrival or moved to the product from a neighboring surface during grinding and packing. The insanitary conditions route therefore collapsed into the same cooking proposition that had decided the first branch and the Butz case before it. With both branches closed, the conclusion followed: failure of the Salmonella performance standard, standing alone, could not support withholding the marks of inspection or withdrawing inspection from Supreme Beef.

The holding was narrow in form and broad in effect, and precision about its scope matters. The court did not strike down the performance standards as measures. The court did not hold that the service lacked authority to sample product, test for Salmonella, or set prevalence expectations for product classes. The court did not hold that an establishment could ignore sanitation requirements or operate a hazard analysis system in name only. The holding addressed the legal consequence the service had attached to one result. A failed standard, by itself, was not adulteration under either paragraph the court examined, and because it was not adulteration, it could not carry the enforcement consequences that an adulteration finding carries, including the withholding of inspection marks and the withdrawal of inspection built on that failure. The distinction between measuring a condition and punishing the measurement became the operative line in meat inspection enforcement after December 2001.

The consequences unfolded along that line. The Food Safety and Inspection Service kept the Salmonella performance standards as verification measures. The service continued to draw samples, test for the organism, publish results, and use establishment level results to verify whether an establishment’s Hazard Analysis and Critical Control Point system was controlling the hazards it was designed to control. A failed standard retained practical force. Failure drew scrutiny, prompted demands for corrective action within the hazard analysis framework, and signaled to the service, to customers such as federal feeding programs, and to the public record that an establishment’s process control lagged behind the level for its class. What failure lost was its stand alone legal force as a ground for closing a plant. After the Fifth Circuit decision, the service could not rest a suspension or withdrawal of inspection on standard failure alone, and any enforcement action of that gravity required a predicate the statute recognized apart from the prevalence count.

Enforcement therefore ran through the channels the decision left open, and the channels were substantial. Sanitation requirements remained directly enforceable. An establishment that failed to maintain sanitary conditions or to follow its Sanitation Standard Operating Procedures could face withholding and suspension action on that basis. Hazard analysis failures remained enforceable: a system that failed to identify a reasonably likely hazard, that set no workable critical limits, or that ignored monitoring deviations presented defects the service could cite and require to be corrected, with escalation for refusal or repetition. Adulterant findings remained the strongest channel of all. Product bearing an organism or substance the service had lawfully declared an adulterant could be retained, denied the marks of inspection, and kept from commerce. Salmonella counts could inform each channel, guiding inspectors toward plants where closer examination was warranted, without themselves constituting the violation cited.

The department said as much in its immediate response. In December 2001, Under Secretary Elsa Murano stated that the Fifth Circuit decision did not change the requirement that plants produce safe product. The statement marked the boundary the department drew around the loss. A particular enforcement theory had been rejected: prevalence of a nonadulterant organism, measured against a performance standard, as a self sufficient ground for withdrawal. The underlying obligations of establishments, to operate under sanitary conditions, to run a functioning hazard analysis system, and to keep adulterated product from commerce, continued on their own statutory footing and did not depend on the rejected theory.

The contrast with Escherichia coli O157:H7 shows that the limit was organism specific and reasoning specific, not a general judicial refusal to let the service treat pathogens as adulterants. On September 29, 1994, Administrator Michael R. Taylor announced that the service would treat Escherichia coli O157:H7 in raw ground beef as an adulterant under the Federal Meat Inspection Act, making contaminated product illegal to sell, and the service began sampling raw ground beef for the organism in October 1994. Industry challenged the policy in Texas Food Industry Association v. Espy. The district court refused to block the policy in 1994, and the Fifth Circuit affirmed in 1995. The same circuit that later rejected the Salmonella enforcement theory therefore sustained the O157:H7 adulterant designation, and the difference between the two outcomes lies in the reasoning each organism’s facts supported under 601(m)(1).

The Espy litigation sustained the designation on grounds the Salmonella record could not supply. Many consumers regard ground beef cooked rare or medium as properly cooked, so ordinary cooking practice does not reliably destroy the O157:H7 organism. The organism also presents a low infectious dose, so small numbers surviving preparation can cause severe illness, including the hemolytic uremic syndrome that had marked the 1993 outbreak traced to undercooked hamburgers. Those facts placed O157:H7 within the paragraph as the court read it: a substance whose presence could render product injurious to health under the cooking practices consumers actually follow. Salmonella failed that test, in the Fifth Circuit’s reasoning, precisely because the cooking proposition ran the other way. Normal cooking destroys Salmonella, and consumers know raw meat requires cooking. One organism’s adulterant status survived and the other’s enforcement theory fell, in the same circuit within the same decade, on the difference between those two cooking records.

Later steps extended the adulterant side of the line without moving the Salmonella side. In 1999 the service extended the O157:H7 adulterant policy to non-intact beef products. In 2002, after a large ground beef recall, ground beef establishments had to reassess their hazard analysis plans for O157:H7. In September 2011 the service announced six additional Shiga toxin producing Escherichia coli strains, O26, O45, O103, O111, O121, and O145, as adulterants in raw non-intact beef, with routine testing beginning June 4, 2012. None of the steps altered the legal position of Salmonella in raw meat and poultry, which remained a nonadulterant organism measured by performance standards whose failure required some further, independently grounded violation before the service could close an establishment.

What the gap means can now be stated without exaggeration in either direction. The service did not lack tools after Supreme Beef. Sanitation enforcement, hazard analysis enforcement, adulterant findings for designated organisms, retention of product, and the verification force of published Salmonella results all remained. The gap was specific and structural. For the bacterial pathogen that the Centers for Disease Control and Prevention burden estimates identify as the leading bacterial cause of foodborne hospitalization and death, responsible on the standing estimate for about 1.35 million infections a year, the direct measure of that pathogen in an establishment’s product could not by itself sustain the service’s gravest sanction. Prevalence evidence had to be translated into another legal language, insanitary conditions shown by recognized means, a hazard analysis system shown to be defective, or an adulterant shown to be present, before withdrawal could follow. Translation also costs time, proof, and litigation exposure that a direct standard would not, and an establishment whose sole demonstrable failing was a prevalence rate far above its class standard occupied ground the statute, as the Fifth Circuit read it, did not let the agency reach on that evidence alone.

The gap is often misdescribed as a judgment about how dangerous Salmonella is. Neither the District of Columbia Circuit in 1974 nor the Fifth Circuit in 2001 held that Salmonella does not cause illness, and the burden figures were not in dispute in either case. The courts reasoned about categories. An added substance, a product rendered injurious, a deleterious change inside an establishment: the definitions in 21 U.S.C. 601(m)(1) and 601(m)(4) sorted organisms and conditions into those the inspection service could condemn directly and those it could measure, publish, and pursue only through other requirements. Salmonella fell on the measured side of the line because it occurs naturally in raw product and because normal cooking destroys it. O157:H7 fell on the condemned side because consumer cooking practice does not reliably destroy it. Until Congress amends the definitions or a court reads them differently, that sorting governs, and the pathogen causing the most measured illness in the food supply remains the pathogen whose prevalence, however extreme in a given establishment’s samples, cannot alone close the plant that produced it.

Recalls, detention, and the tools each agency actually holds

The most consequential limit on the Food Safety and Inspection Service is a power the service never received. The Federal Meat Inspection Act, enacted on June 30, 1906, and substantially amended by the Wholesome Meat Act of 1967, authorized inspection before and after slaughter. The Poultry Products Inspection Act of August 28, 1957, and the Egg Products Inspection Act of December 29, 1970, extended a parallel structure to poultry and to liquid, frozen, and dried egg products. None of those statutes, in the form in force on February 1, 2021, authorized the Secretary of Agriculture to order an establishment to remove product from commerce. Because a direct order was unavailable, the service built a practice of requested recalls and a set of adjacent tools that acted on product inside a plant, product in commerce, imports at the border, and the inspection status on which an establishment depended for lawful operation. Together, those tools supplied substantial leverage, but none of them was a recall order, and the distance between leverage and authority became the central dispute in repeated congressional debate.

Can the Secretary order a meat plant to recall its product?

No. Under the meat, poultry, and egg statutes in force on February 1, 2021, the Secretary of Agriculture had no power to order a recall. A recall remained the establishment’s voluntary decision, often made after a request from the Food Safety and Inspection Service, which then classified, publicized, and verified the action.

The voluntary character of a recall followed from the text and structure of the inspection statutes. The meat act defined adulterated and misbranded product, prohibited commerce in such product, prescribed inspection, and authorized detention, seizure, and condemnation through specified procedures, and the poultry and egg acts followed the same pattern. A recall, meaning a firm’s removal or correction of product that had left the firm’s direct control and entered distribution, appeared in agency practice rather than as a commanded remedy in the statute. The establishment owned the decision to initiate a recall and bore the work of retrieving product through its distribution records and customer contacts. That division of labor could produce a fast and broad removal when a firm cooperated at once, and it could also produce negotiation over whether a recall was warranted, over scope, and over timing, while product remained available to consumers.

A recall ordinarily began before any public announcement, with a preliminary inquiry into whether product in commerce presented a problem that justified removal. Information could reach the service through inspection personnel in an establishment, through laboratory results from agency sampling, through reports of illness, or through a firm’s own notification that product might be adulterated or misbranded. The product had to be identified with enough precision to separate it from unaffected production, and the problem had to be stated in terms the statutes recognized, such as the presence of an adulterant or a misbranding failure. The period of production, the amount produced, and the places to which product had been shipped all mattered, because a recall that could not describe its own scope could not be carried out or verified.

When the preliminary facts pointed toward a recall, the matter moved to a recall committee process, described here in the general terms that agency practice supported on February 1, 2021. The committee brought together program areas with different knowledge of the case, including inspection, technical and laboratory, and public communications functions, to evaluate the evidence and recommend whether the establishment should be asked to recall product and, if so, what product and with what public notice. The recommendation that emerged was a request to the firm, not an order, and the firm could accept it, propose a different scope supported by records, or decline to recall. Decline remained legally possible even after a committee recommendation, although a firm that declined faced the adjacent tools described below and the publicity that accompanied agency action. The committee process supplied deliberation and internal consistency, but it did not supply compulsion.

Classification translated the health assessment into a public signal, as the service classified recalls into Class I, Class II, and Class III according to the health risk presented. A Class I recall involved a health hazard situation in which a reasonable probability existed that use of the product would cause serious adverse health consequences. A Class II recall involved a situation in which a probability of adverse health consequences existed but was more limited or more remote, or in which the consequences, if they occurred, would be less severe. A Class III recall involved a situation in which use of the product would not be expected to cause adverse health consequences, although the product was nevertheless subject to recall, for example because of a labeling failure that required removal. Retailers, distributors, and consumers used the classification to judge the urgency of shelf checks and to distinguish a recall driven by a serious hazard from one driven by a defect that carried no expected illness.

Public notification and effectiveness checks completed the sequence. When an establishment conducted a recall, the service publicized it so that notice reached beyond the firm’s immediate customers, identifying the establishment, the product, the reason for the recall, and the classification in terms that allowed holders to recognize the product. After notice, agency personnel conducted effectiveness checks, contacting or visiting consignees and other holders, on a basis proportioned to the recall, to verify that notice had been received and that product had been removed or controlled. Gaps in the firm’s distribution list, consignees that had not acted, and product found in commerce after notice all showed whether the voluntary effort was working. Where checks showed failure, the service could press the firm to expand or repeat notification and could turn to detention or other tools for product within reach, but the remedy for a weak recall remained pressure and publicity rather than conversion of the recall itself into an agency order.

Administrative detention was the first of those separate tools, and the one most often paired with a recall request. Under 21 U.S.C. 672, an authorized representative of the Secretary could detain meat or a meat food product found in commerce for a period not to exceed 20 days, pending action under the seizure provisions or notification of an appropriate authority, when the representative had reason to believe the product was adulterated or misbranded or had not been inspected and passed, and parallel detention provisions applied under the poultry and egg products statutes. A detained lot could not move as ordinary commerce while detention held, which gave the service time to complete testing, to assemble the factual basis for seizure, or to secure the firm’s agreement to a recall or to voluntary destruction where the law allowed it. The 20 day limit defined the tool as a hold rather than a removal, because detention did not condemn product, did not retrieve product already sold to consumers, and did not reach product the agency could not locate. Its strength was immediacy where product sat in quantity at a warehouse or distributor, and its weakness was duration and geography.

Judicial seizure and condemnation supplied the compulsory removal that detention only prepared. When the Secretary sought permanent control of adulterated or misbranded product in commerce, the meat act provided for a proceeding against the product itself, brought through the United States Attorney within the Department of Justice, in which a court could condemn the product and order its disposition, and parallel provisions governed poultry and egg products. A court, rather than the agency or the firm, decided whether the statutory basis for condemnation had been established, and seizure took the product into legal custody so that commerce in it stopped while the case proceeded. Condemnation could end in destruction or in another disposition the court ordered, and the route did not depend on the firm’s consent. The route stood behind detention and recall requests in ordinary practice because a proceeding took preparation and judicial time and addressed only the lots named, rather than the whole distribution pattern a cooperative firm could sweep through at once. Seizure was a real compulsory power, but a lot by lot judicial power, ill matched on its own to the speed and breadth expected of a large recall.

Retention operated earlier in the chain, inside the establishment, before product entered commerce. Inspection personnel could retain product within a plant when the product appeared adulterated or otherwise not eligible to move, holding it under agency control while the establishment and the service resolved its status. Retained product stayed where inspectors could see it and where plant movement could be stopped without a search through distribution, and product that could lawfully be brought into compliance through a permitted step could be released after that step and after agency verification. Product that could not be brought into compliance could be condemned or destroyed under applicable procedures rather than shipped. Retention illustrated a general feature of the inspection statutes, because control was strongest at the point of production, where inspection presence placed agency personnel beside the product and the marks of inspection controlled release into commerce, and control weakened as product moved outward.

Withholding and suspension of inspection attacked a different pressure point, the establishment’s legal ability to operate at all. Under the meat act, antemortem and postmortem inspection were conditions of lawful slaughter, an establishment could not lawfully operate slaughter without inspection coverage, and inspection presence for processing establishments under the standard model came at least once per operating day. The Secretary could withhold or suspend inspection in circumstances the statutes and regulations specified, including certain egregious conditions for which suspension without a prior hearing was available, subject to notice and to procedures that gave the establishment an opportunity to respond. Suspension did not order any consumer or distributor to return product, and its effect was more basic, because a plant from which inspection was suspended could not slaughter under inspection and could not apply the marks of inspection to product, so the ordinary lawful business of the plant stopped for the period of suspension. Suspension was therefore a powerful lever for obtaining cooperation, including cooperation with a recall request, but it remained an indirect lever for product already in homes and stores, because closing a plant did not retrieve a single package from a retail shelf.

Refusal of entry completed the set of tools at the border. Imported meat, poultry, and egg products entered the United States only from foreign inspection systems that the service had determined to be equivalent to the United States system, and imported product was subject to reinspection at import establishments. Product that failed reinspection, that came from a system or establishment not eligible to ship, or that otherwise failed the conditions of admission could be refused entry, and refused product never acquired the lawful status of product in domestic commerce. The agency acted at a controlled point, on identified shipments, before distribution dispersed the product, which made refusal cleaner than any remedy applied after admission. The limit was equally clear, because refusal governed only the shipment presented at the border and lost its advantage after imported product had been admitted and distributed, when the service stood in the same position as with domestic product.

The contrast across the dinner plate made the gap visible in a single grocery purchase, because a frozen pepperoni pizza fell under the inspection service on account of its meat topping, while a frozen cheese pizza fell under the Food and Drug Administration. After the FDA Food Safety Modernization Act, Public Law 111-353, signed on January 4, 2011, the legal consequences of that division included a difference in recall authority. Section 206 of the 2011 act gave the Food and Drug Administration mandatory recall authority over food within its jurisdiction, other than infant formula, which Congress addressed under a separate provision of the food statute. The agency could exercise the authority where the agency determined that a reasonable probability existed that the food was adulterated or misbranded in the relevant statutory sense and that use of or exposure to the food would cause serious adverse health consequences. Before an order issued, the responsible party received an opportunity for an informal hearing, at which the party could present views on whether a recall was warranted. If the party refused to recall voluntarily, the Food and Drug Administration could order the recall without the firm’s consent.

The Food and Drug Administration also held an older administrative detention power for food, and the two detention regimes showed that detention and recall were distinct instruments in both systems. Under section 304(h) of the Federal Food, Drug, and Cosmetic Act, as expanded by the Bioterrorism Act of 2002 and by the 2011 food safety act, the Food and Drug Administration could order detention of food for a limited period when an officer or qualified employee had credible evidence or information indicating that the food presented a threat of serious adverse health consequences. Detention held the food in place while the agency decided whether to pursue seizure or another action, much as detention under 21 U.S.C. 672 held meat in place while the inspection service decided its next step. Congress had given both agencies a temporary hold and had given the courts a seizure remedy reachable under both systems, yet in 2011 Congress gave a recall order to one agency and not to the other.

Label approval added a quieter lever that reinforced the plant level side of that answer. For meat, poultry, and egg products, labels required prior approval by the inspection service, subject to the expansion of generic approval over time for labels that fit established categories, and a label that had not received a required approval could not lawfully be used. Prior approval gave the service a checkpoint before product reached commerce, at which claims, ingredient statements, handling instructions, and other label elements passed agency review. The Food and Drug Administration’s food labeling regime operated largely after the fact, because a food label within that jurisdiction generally reached the market without prior federal approval of the label itself, and the agency acted against false or misleading labeling through post-market enforcement tools. The difference cut in the opposite direction from the recall difference, since on labeling the inspection service held the earlier and more systematic checkpoint, while on recalls the Food and Drug Administration held the direct order.

Congress considered closing that gap directly and, through February 1, 2021, never did. Bills to grant the Secretary of Agriculture mandatory recall authority over meat, poultry, and egg products were introduced repeatedly over the years and were not enacted, and the recurring design, stated in general terms, would have authorized the Secretary to order a recall where product presented a defined health risk, with notice to the establishment and an opportunity for a hearing or another form of review. The failure of the proposals cannot be explained by absence of a triggering record, because the 1993 outbreak of Escherichia coli O157:H7 infection traced to undercooked ground beef, in which the Centers for Disease Control and Prevention’s retrospective account counted more than 700 illnesses in four states and four children killed, had already shown how widely a contaminated ground beef product could travel before removal. Enactment nevertheless required agreement that a new order would improve on the combination of requested recalls, detention, seizure, and suspension that the statutes already supplied, and that agreement never formed.

The case for a mandatory order, as consumer advocates presented it, rested on parity, speed, and accountability. Parity came first after 2011, because the Food and Drug Administration could order recalls of most food while the Secretary of Agriculture could only request recalls of meat, poultry, and egg products presenting a comparable risk. Advocates argued that protection should not depend on whether a product contained enough meat to cross a jurisdictional threshold stated in agency practice, commonly described as more than 3 percent raw meat or 2 percent cooked meat or poultry. Speed followed, because a firm deciding whether to volunteer could weigh reputation, cost, and legal exposure while product remained on sale, and negotiation over scope consumed time that a direct order would not consume in the same way. Accountability completed the case, because a voluntary act left the public record centered on the firm’s announcement even when agency pressure had produced it, and advocates answered the claim that voluntary recalls almost always occurred by pointing to the cases that mattered most, in which a firm hesitated, disputed the evidence, or proposed a scope narrower than the distribution records supported.

The case against a mandatory order, as industry and department witnesses presented it, rested on performance, leverage, and the risk that a formal order would slow the removal it aimed to accelerate. Performance came first, because voluntary recalls, as witnesses testified across the repeated consideration of the bills, almost always occurred when the service asked, so a mandatory power would address a refusal the record showed to be exceptional. Leverage followed, because the service did not approach a firm empty handed, since detention could hold product in commerce for up to 20 days, seizure could take product through a court, retention could hold product in the plant, refusal of entry could stop imports, and suspension could stop the plant itself. Publicity added force to each tool, because a public agency statement that product presented a Class I hazard damaged the same reputation a firm protected by cooperating. The final argument concerned process, because a recall order that carried hearing rights and judicial exposure would invite contest at the moment of decision, while a request could be made on the evidence available at once, so the informality of the voluntary system was itself a source of speed.

Both presentations contained a point the other side had difficulty displacing. The advocates were correct that leverage, however strong, was not authority, and that every compensating tool had a boundary a recall order would not have had, since detention expired after 20 days, seizure proceeded lot by lot through a court, retention stopped at the plant gate, suspension stopped a plant rather than retrieving product, and refusal of entry stopped at admission. The witnesses were correct that the boundaries seldom governed the outcome, because the combination of tools, the economics of a plant dependent on inspection, and the reputational cost of resisting a Class I public notice made cooperation the rational course for nearly every firm in nearly every case. Through February 1, 2021, Congress left the answer in the older statutes, under which the Secretary could detain, seize through a court, retain, refuse, withhold, and suspend, and could ask, classify, publicize, and verify. The one act the Secretary could not perform was the act that gave the process its name, and a recall of meat, poultry, or egg products remained, in law, the establishment’s own.

How a rule gets written, and why the system stays split

A statute that orders continuous inspection and forbids slaughter without an inspector present does not operate itself. Between the text Congress enacted and the decisions made on a slaughter line stands a second body of law, produced not by Congress but by the Department of Agriculture through the Food Safety and Inspection Service. That body consists of regulations published after public proceedings, of economic and risk analyses prepared to justify those regulations, and of a quieter layer of guidance, directives, and notices that translates regulatory language into instructions an inspector follows on a particular shift. The machinery defines what the inspection service can change on its own authority and what it cannot, and it frames a durable puzzle. If the division of responsibility between the Department of Agriculture and the Food and Drug Administration produces lines that follow toppings, sandwich construction, and the presence or absence of an eggshell, why has that division survived decades of findings that it fragments federal oversight? The answer lies less in any defense of the lines than in the statutes and committee jurisdictions that hold them in place.

The Administrative Procedure Act path from proposal to final rule

The ordinary road for a binding FSIS regulation runs through the Administrative Procedure Act. The Secretary of Agriculture, acting through the inspection service, publishes a proposed rule in the Federal Register, states the basis and purpose of the proposal, and opens a period for public comment. Establishments that would bear the cost of a new requirement, consumer organizations pressing for a stricter standard, state officials, trade associations, and individual inspectors file into the same docket, and the agency must confront that record when it acts. Agency practice posts rulemaking documents through Regulations.gov, replacing a historically paper based docket system, so that a proposal, its analysis, and the comments answering it sit in a public file a court or a regulated establishment can later inspect.

A final rule published in the Federal Register responds to the significant comments received, explains departures from the proposal, and states an effective date and, where retooling a plant process requires time, compliance dates set out in the rule itself. That discipline does substantive work, because many FSIS rules turn on contested technical judgments: whether a pathogen should be treated as an adulterant, whether a sampling frequency verifies a process or merely measures incoming material, whether a sorting task can move from a federal inspector to a trained establishment employee without weakening the carcass by carcass determination the statutes require. A court reads the preamble against the comment record, and an agency that has ignored a substantial objection invites reversal or remand. The procedure also sets a slow tempo: a major rule must be drafted, cleared inside the department, reviewed outside it when significant, published for comment, reconsidered, cleared again, and published in final form. Each stage protects participation, reasoned decision, and fair notice, and each gives opponents a point at which to press an objection or prepare litigation. The status quo therefore holds the advantage of already being codified.

Who reviews USDA rules before they are issued?

The Secretary of Agriculture issues Food Safety and Inspection Service rules after proposal and comment under the Administrative Procedure Act, with significant rules reviewed by the Office of Information and Regulatory Affairs and major rules also reviewed within the Department of Agriculture by the Office of the Chief Economist and its risk analysis office.

The review structure has two layers, one in the Executive Office of the President and one inside the Department of Agriculture, and a major rule must satisfy both. The external layer comes from Executive Order 12866, issued in 1993, which established centralized regulatory review for executive agencies. Under that order, the Office of Information and Regulatory Affairs, housed in the Office of Management and Budget, reviews significant regulatory actions before publication, testing whether the agency has identified a problem meriting federal action, considered alternatives, and assessed benefits and costs under the principles the order prescribes. A proposal to restructure slaughter inspection or impose a new performance standard therefore does not move on food safety judgment alone: officials with no institutional stake in meat inspection read the draft and hold clearance until the analysis answers objections that may concern economy wide regulatory policy as much as microbiology.

The internal layer rests on statute. The Department of Agriculture Reorganization Act of 1994 created the Office of Risk Assessment and Cost-Benefit Analysis within the Office of the Chief Economist to review regulatory impact analyses and risk analyses for major rules across the department, including rules of the inspection service. That office examines whether a risk assessment separates the probability of contamination from the severity of illness in a defensible way, whether baseline illness estimates rest on named sources, whether projected reductions follow from the intervention proposed, and whether compliance costs for large, small, and very small establishments reflect their different capacities to absorb new testing, recordkeeping, or staffing expenses. Cost-benefit analysis requirements trace to Executive Order 12866 and, for the department, to the 1994 reorganization act, and together cost-benefit and risk analysis form the grammar a major rule must speak. A claim that a pathogen reduction step is desirable states a value. A claim that a specified sampling regime, at specified frequencies in specified classes of establishments, will avert an estimated range of illnesses at an estimated cost, with uncertainties stated and alternatives compared, states a case in the form reviewers are instructed to credit. That form disciplines agencies against imposing expensive obligations for trivial gains and exposes assumptions a purely legal justification could leave buried. It also carries a structural tilt. Costs arrive as engineering estimates that lend themselves to quantification, while benefits depend on epidemiological estimates of illnesses that never occur because a process changed, illnesses that leave no case report to count. Reviewers under Executive Order 12866 are directed to recognize benefits that are difficult to quantify, but the asymmetry remains.

Guidance, directives, and notices as the operating layer

Below regulations sits the layer through which much of the operating system actually reaches plants. The inspection service issues directives instructing its personnel on inspection methods, verification procedures, and enforcement protocols, issues notices announcing time limited instructions or recalling attention to procedures, and publishes guidance telling establishments how the agency understands a regulatory obligation and what compliance approaches it will accept. None of these instruments carries the formal status of a legislative rule adopted through notice and comment, a distinction that matters when an establishment challenges an agency action in court. In the daily life of a plant, however, content matters more than status. An inspector verifying sanitation performance or deciding whether a deviation has been corrected acts on the directive governing that task, not on an unaided reading of the Federal Meat Inspection Act.

The sub rule layer exists because regulations cannot carry the full weight of continuous inspection without becoming unusable. A regulation states that an establishment must maintain Sanitation Standard Operating Procedures, conduct a hazard analysis, and operate under a Hazard Analysis and Critical Control Point plan. It cannot sensibly specify, in codified text amended only through full rulemaking, every verification frequency, form, and sampling instruction across thousands of establishments. Directives and notices fill that space and can be revised at a pace regulations cannot match when a sampling pattern changes. The same flexibility creates a standing tension. Guidance can function in application as the standard an establishment must meet to avoid noncompliance findings, while escaping the comment process and centralized review a regulation would face. The practical settlement is visible in the plant: the statute supplies the authority, the regulation supplies the enforceable program, and the directive supplies the working instructions that determine what an inspector asks for on a given morning.

The modernization rules as the machine at work

The modernization rules show the machinery at full stretch. In 2014 the inspection service finalized a modernization of poultry slaughter inspection, establishing the New Poultry Inspection System. On October 1, 2019 the agency finalized the New Swine Slaughter Inspection System, the first change to swine slaughter inspection in decades. Both systems were optional: an establishment could elect the new system or remain under the traditional arrangement. Under both systems, some sorting tasks shifted to establishment employees, who would identify and remove carcasses or parts presenting obvious defects before federal inspection, while federal inspectors retained the carcass by carcass determination on which the marks of inspection depend. That distinction carried the legal weight of the rules. Sorting could be characterized as a preliminary production function establishments already performed for quality purposes. Determination, the official judgment that a carcass is not adulterated and may pass, remained a governmental act the statutes reserve to federal personnel. Each rule drew extensive comment, from operators supporting line flexibility and from inspector organizations and consumer organizations questioning whether shifting sorting to plant employees weakened the independence of the first look at each carcass, and each supporting analysis had to estimate effects on staffing, labor, line operations, and food safety outcomes in the terms the departmental and presidential reviewers require. The governing limit is plain: the agency can modernize the deployment of its inspectors by regulation, but it cannot by regulation repeal the continuous inspection obligation, transfer the official determination to private employees, or rewrite adulteration definitions litigation has already construed. Those limits belong to Congress.

The egg products rule completed the demonstration from a different direction. On October 29, 2020 the inspection service published the Egg Products Inspection Regulations final rule at 85 FR 68640, requiring egg products plants to develop Sanitation Standard Operating Procedures and to operate under Hazard Analysis and Critical Control Point systems, the two pillars imposed on meat and poultry establishments by the rulemaking of the nineteen nineties. Egg products, meaning eggs removed from the shell and processed as liquid, frozen, or dried product, had long received continuous inspection under the Egg Products Inspection Act of 1970, but continuous presence had not by itself produced the establishment owned systems the meat and poultry rules made mandatory. The 2020 rule closed that gap through the standard machinery, with compliance dates the rule itself prescribed in late 2021 and in 2022, giving plants a scheduled interval to build sanitation procedures, conduct hazard analyses, and train personnel. By February 1, 2021 the direction of regulation across all three inspection statutes pointed the same way, toward establishment systems verified by federal personnel and justified in the language of risk and cost, even as the oldest questions about federal food safety structure remained exactly where Congress had left them.

The fragmentation finding and the single agency proposal

Against that machinery stands a counter reading, the most persistent external judgment on the system for more than four decades. In the repeated formulation of the Government Accountability Office, federal food safety oversight is fragmented across 15 federal agencies administering more than 30 statutes. The inspection service and the Food and Drug Administration form the largest part of that structure, with the food and drug agency covering roughly four fifths of the United States food supply and the inspection service roughly one fifth, on the GAO formulation in GAO-17-74, while close to the same food safety funding supports each share. In 2007 the GAO placed federal food safety oversight on its High-Risk List, and the designation remained part of the oversight landscape through February 1, 2021, on the allegation that the structure assigning the tasks prevented a unified, risk based strategy across the whole food supply.

The structural recommendation behind the finding has an equally long record. The Government Accountability Office has recommended a single, independent food safety agency in various forms since the late 1970s, returning to the proposal repeatedly as administrations and Congresses declined to enact it. In GAO-17-74 in 2016 the office recommended a government wide national strategy rather than resting its case only on the structural call, a shift acknowledging how often that call had been made and refused. The legislative record matches the pattern. The Safe Food Act, introduced repeatedly from 2005 onward by Senator Durbin and Representative DeLauro, would have created a single Food Safety Administration consolidating functions statutes had scattered across departments. A 2015 presidential budget proposal sought a single food safety agency within the Department of Health and Human Services, moving inspection functions into the department housing the Food and Drug Administration. None of these vehicles was enacted. Supporters of consolidation, including the Government Accountability Office in its structural reports and the sponsors of the Safe Food Act, built the affirmative case on coherence: a single agency could set priorities by risk rather than by whether a product contains a threshold amount of meat, apply one inspection philosophy calibrated to hazard instead of two divided by commodity, present one accountable official when a multi ingredient product crossed jurisdictional lines during an outbreak investigation, and end duplication between two field forces and regulatory staffs serving adjacent segments of the same supply. The recurring examples state the case in miniature: frozen cheese pizza falls to the Food and Drug Administration while frozen pepperoni pizza in the same plant falls to the inspection service. No risk assessment written from first principles would draw that line, and a line no risk assessment would draw, on the consolidation argument, should not govern the allocation of federal inspection resources.

Coordination without merger, and why it has not settled the question

Between the inherited structure and a single agency lies the course administrations actually took: coordination short of merger. In 2009 the Food Safety Working Group was established to coordinate federal food safety efforts across the departments holding food safety responsibilities, bringing agriculture and health officials into a common forum charged with aligning priorities and identifying gaps no single agency could close alone. Interagency agreements supplied the working instruments, allocating responsibilities at shared borders, establishing procedures for joint action during outbreaks involving products under divided jurisdiction, and supporting joint initiatives on produce and on Salmonella. Such instruments operate without amending a section of the inspection acts or of the Federal Food, Drug, and Cosmetic Act, and coordination has real value wherever friction, rather than absent authority, is the obstacle. Its limit is structural. Agreements rest on the continued consent of the agencies that signed them, bind no future administration, appropriate no funds, and transfer no authority, so each cross border outbreak tests arrangements that personnel turnover and competing budgets erode. By February 1, 2021 coordination had neither failed conspicuously enough to force consolidation nor succeeded completely enough to retire the proposal.

The opposition case at full strength

The opposition begins with Congress, because a transfer of functions is also a transfer of jurisdiction, and committees defend jurisdiction as the working capital of legislative influence. The agriculture committees authorize and oversee Department of Agriculture programs, and agriculture appropriations subcommittees fund the inspection service. Moving inspection into the Department of Health and Human Services, or into a new independent agency, strips those committees of jurisdiction over one of the largest programs in their portfolio and delivers it to committees organized around health policy. Stated at full strength, the argument treats jurisdiction as accumulated competence rather than turf: agriculture committees hold the hearing record on inspection, maintain relationships with state programs operating at least equal to the federal program in roughly half the states, and answer to constituencies in which slaughter and processing establishments are major employers.

The second argument is statutory, and the strongest on purely legal grounds. The inspection acts impose a mandatory, continuous regime: antemortem examination of livestock, postmortem inspection of every carcass and its parts, inspector presence during operating hours in slaughter establishments, and inspection presence at least once per operating day in processing establishments under the standard model. The Federal Food, Drug, and Cosmetic Act imposes a different regime, built on facility registration, preventive controls, and periodic inspection at frequencies the FDA Food Safety Modernization Act set in statute, with high risk domestic facilities inspected at least once every three years and others at least once every five years. These are not two administrative styles a merger memorandum could harmonize. They are two statutory commands specifying different frequencies, triggers, and relationships between inspector and establishment. A consolidated agency would inherit both commands on its first day and would inspect adjacent plants on adjacent schedules until Congress amended one regime or the other, at which point the consolidation debate would reopen inside the new agency as a contest over which model should prevail. Merger changes the organization chart faster than it changes the law the organization administers.

The third argument concerns the workforce. Continuous inspection is delivered by thousands of inspectors, veterinarians, and support personnel stationed in and around establishments across the country, and that daily presence is the system working asset rather than overhead attached to it. Figures recorded by the Congressional Research Service for the period around 2003, about 7,700 field employees including roughly 1,000 veterinarians at some 6,200 plants and import stations, convey the order of magnitude a reorganization must move, merge, or reassign. A consolidation that relocates reporting lines, merges personnel systems, or restations field staff risks attrition among the experienced inspectors on whose judgment the carcass by carcass determination depends, across a transition measured in years, throughout which slaughter establishments cannot lawfully operate without inspection coverage. The benefits of consolidation are projected and long term; the disruption costs are immediate and fall on the one function that cannot be suspended while a transition completes. The fourth argument denies that structure is the binding constraint at all. On that view, slow information sharing during outbreaks, inconsistent sampling priorities, duplicated data systems, and gaps at jurisdictional borders are management failures, and management failures travel with the managers into any structure inheriting the same personnel, appropriations process, and statutory mandates. A single administrator handed two incompatible inspection statutes, two field cultures, and a budget assembled from two appropriations streams faces the same coordination problem two administrators face. The remedies matching that diagnosis, common data standards, funded coordination, and joint strategic planning of the kind GAO-17-74 recommended in 2016, are available without the transition risk a merger guarantees. The opposition case, at full strength, does not defend topping based lines as rational. It argues that rationalizing them by merger purchases an uncertain gain at a certain cost, under statutes a merger alone does not amend.

The split therefore persists because the Federal Meat Inspection Act of 1906, the Poultry Products Inspection Act of 1957, and the Egg Products Inspection Act of 1970 assign continuous inspection to the Department of Agriculture in statutory text only Congress can move, because the Federal Food, Drug, and Cosmetic Act of 1938 assigns the remainder of the food supply to the Food and Drug Administration in equally durable text, because the committees that would have to report a consolidation bill are the committees consolidation would diminish, and because appropriations, state cooperative programs, labor structures, and plant routines built on those assignments renew the arrangement every operating day. Rulemaking operates entirely inside that settlement. Through the Administrative Procedure Act, through review by the Office of Information and Regulatory Affairs and the Office of Risk Assessment and Cost-Benefit Analysis, and through the directives that carry regulations into plants, the inspection service modernized poultry and swine inspection and extended sanitation and hazard analysis systems to egg products, all without altering by one product the jurisdictional lines that make USDA food safety a partial system within a fragmented whole. The lines hold because statutes and committee jurisdictions hold them, not because anyone designed the arrangement as a system, and they will move only when Congress writes a different structure into law.

The jurisdiction and authority table

Agency	Statutes administered	Products covered	Inspection model	Recall power	Enforcement standard
Food Safety and Inspection Service	Federal Meat Inspection Act, Poultry Products Inspection Act, and Egg Products Inspection Act	Meat, poultry, liquid, frozen, and dried egg products, and Siluriformes fish	Continuous inspection, with examination before slaughter and of every carcass after slaughter, and daily presence in processing establishments	No mandatory recall authority; recalls are voluntary establishment actions, often at agency request	Adulteration findings, sanitation requirements, and hazard analysis system failures, enforced through retention, detention, and withholding or suspension of inspection
Food and Drug Administration	Federal Food, Drug, and Cosmetic Act, as amended by the FDA Food Safety Modernization Act	Most other food, including shell eggs, seafood other than Siluriformes fish, produce, and packaged foods without sufficient meat or poultry content	Risk based periodic inspection of food facilities, with high risk domestic facilities inspected at least once every three years and others at least once every five years	Mandatory recall authority over food other than infant formula, after an opportunity for an informal hearing, where a reasonable probability of serious adverse health consequences exists	Adulteration and misbranding provisions of the general food statute, enforced through detention, seizure, and, after the 2011 act, recall orders

Authority inherited and authority limited

Statutory inheritance decided which agency held which food, and one word of statutory text, adulterated, construed by a court against the facts of ordinary cooking, decided how far the strongest enforcement power of the Agriculture Department agency could reach, so the gap in the system is a gap between authority as inherited and regulation as delivered. Congress assigned meat in 1906, poultry in 1957, and egg products in 1970, and later Congresses amended each statute inside its own track rather than redrawing the boundary between tracks. Inside that inherited boundary, the Fifth Circuit reading of rendered in Supreme Beef Processors left Salmonella prevalence measurable but not, standing alone, punishable by withdrawal of inspection. Inheritance supplied the map. Judicial construction supplied the limit.

The modernization record shows how much change could occur inside those limits and how much could not. The FDA Food Safety Modernization Act, signed on January 4, 2011, rebuilt the preventive framework on the Food and Drug Administration side of the line, adding mandatory recall authority and statutory inspection frequencies, and the Food Safety Modernization Act prevention framework shows a legislature willing to grant compulsion to one agency while leaving the parallel request based recall practice untouched on the other side. Agricultural legislation moved individual assignments without disturbing the structure, as the farm act of 2008 transferred Siluriformes fish to the inspection service, a reminder that the history and structure of the farm bill is also part of the history of food safety jurisdiction. Rulemaking modernized poultry inspection in 2014, swine inspection in 2019, and egg products sanitation and hazard systems in a final rule of October 29, 2020, and none of those steps altered the adulteration definitions or the recall gap.

Labeling and certification show the same pattern of adjacent systems operating by different rules. The inspection service approved labels for meat, poultry, and egg products before sale, with generic approval expanded over time, while Food and Drug Administration food labeling operated largely after placement in commerce, a contrast in machinery that the FDA approval mechanics guide helps place beside the approval habits of the drug side of that agency. Certification regimes added a further layer, because a federal organic claim on a label rested on a separate production statute and its own standards, traced in the history of the federal organic production statute, rather than on the inspection statutes that governed safety in the same plant. A single package could therefore carry a safety legend, an approved label, and a certification claim governed by a third legal regime.

Reading the system as law requires reading several statutes at once and reading the decisive words closely. The inspection acts, the general food statute, and the detention and seizure provisions each supply one part of the operative rule, and a method for how to read a federal statute as enacted text, definition by definition and section by section, is the method the subject demands. Supreme Beef Processors turned on exactly that discipline. Added, injurious, and rendered carried the outcome, and no appeal to the scale of Salmonella illness, estimated by the Centers for Disease Control and Prevention at about 1.35 million infections a year, could move an organism across a definition that the court read the organism to fall outside.

The assessment that follows from the record is neither that continuous inspection failed nor that the divided system lacked capable tools. Detention, seizure, retention, suspension, label approval, equivalence for imports, and hazard analysis verification gave the inspection service real leverage, and establishments almost always recalled product when the service asked. Leverage, however, is not authority, and a system that depends on leverage leaves its hardest cases, a hesitant firm, a disputed scope, an extreme Salmonella prevalence with no independent violation, to negotiation and publicity rather than to a direct legal command. Until Congress amends the definitions in sections 601 and following of title 21, grants a comparable recall power, or consolidates the structure against the objections that defeated consolidation through February 1, 2021, the gap remains a creature of enacted text. For study of that text and the cases construing it, readers can keep your statute notes, citations, and case chronologies together free on VaultBook and can practice and revise US government and civics material on ReportMedic while tracing how inheritance and one defined word produced the system in operation.

Frequently Asked Questions

Q: Which foods does USDA inspect instead of FDA?

The Food Safety and Inspection Service, part of the Department of Agriculture, inspects meat, poultry, and processed egg products under three statutes: the Federal Meat Inspection Act of 1906, the Poultry Products Inspection Act of 1957, and the Egg Products Inspection Act of 1970. Processed egg products means eggs removed from the shell and prepared as liquid, frozen, or dried products. The service also inspects Siluriformes fish, including catfish, transferred from the Food and Drug Administration by the farm act of 2008. The Food and Drug Administration covers most other food, including shell eggs, other seafood, produce, and packaged foods that contain too little meat or poultry to fall under the inspection statutes. Jurisdiction therefore follows the product class named in a statute, rather than a general judgment about which food presents the greater risk.

Q: Why is a cheese pizza regulated differently from pepperoni by USDA?

A frozen cheese pizza is regulated by the Food and Drug Administration under the Federal Food, Drug, and Cosmetic Act, while a frozen pepperoni pizza is regulated by the Food Safety and Inspection Service under the Federal Meat Inspection Act. The difference is the meat topping. Pepperoni brings the finished product inside the meat inspection system, while crust, sauce, and cheese alone do not. Both products can be made in the same plant, with each production run answering to a different agency, a different inspection rhythm, and a different label pathway. The line follows ingredient and product category under the statutes, rather than a comparison of risk between the two pizzas. In agency practice, products containing more than small threshold amounts of meat or poultry, commonly stated as more than 3 percent raw meat or 2 percent cooked meat or poultry, are treated as inspection service products.

Q: What is continuous inspection at USDA?

Continuous inspection is the model required by the meat, poultry, and egg products statutes administered by the Food Safety and Inspection Service. Inspectors examine livestock before slaughter and examine every carcass and its parts after slaughter, and inspectors are present on the line during all operating hours of a slaughter establishment. An establishment cannot lawfully slaughter without inspection coverage. Processing establishments receive inspection presence at least once per operating day in the standard model, a pattern the Congressional Research Service has described as treated as continuous inspection in keeping with the inspection laws. Egg products plants operate under continuous inspection as well. The model differs from the Food and Drug Administration approach for food facilities, which relies on periodic, risk based inspection rather than attendance during production.

Q: How did an outbreak change USDA rules?

In January 1993, an outbreak of Escherichia coli O157:H7 infection traced to undercooked ground beef hamburgers served by the Jack in the Box chain caused, in the retrospective account of the Centers for Disease Control and Prevention, more than 700 illnesses across four states, and four children were killed. On September 29, 1994, Administrator Michael R. Taylor announced that the Food Safety and Inspection Service would treat the organism as an adulterant in raw ground beef, and sampling began in October 1994. In February 1995 the service proposed a Pathogen Reduction and Hazard Analysis and Critical Control Point rule, and the final rule issued on July 25, 1996, added Sanitation Standard Operating Procedures, generic Escherichia coli process control testing, Salmonella performance standards, and mandatory hazard analysis systems phased in by establishment size between 1997 and 2000.

Q: Why can USDA not enforce Salmonella standards directly?

In Supreme Beef Processors, Inc. v. United States Department of Agriculture, 275 F.3d 432, decided in December 2001, the Fifth Circuit held that failure of a Salmonella performance standard, standing alone, could not support withholding the marks of inspection or withdrawing inspection. Under 21 U.S.C. 601(m)(1), Salmonella is not an added substance, because the organism occurs naturally in a substantial proportion of raw meat and poultry, and its presence does not render product injurious to health because normal cooking destroys the organism. Under 21 U.S.C. 601(m)(4), the word rendered requires a deleterious change during preparation, packing, or holding inside the establishment, and the Salmonella level of incoming raw material is a characteristic of that material rather than such a change. The standards remained in use as verification measures, but closure required another statutory ground.

Q: Is E coli an adulterant in USDA regulated beef?

Escherichia coli O157:H7 is treated as an adulterant in raw ground beef under the Federal Meat Inspection Act, under a policy announced on September 29, 1994, by Administrator Michael R. Taylor of the Food Safety and Inspection Service. Product bearing the organism is therefore illegal to sell, and the service began sampling raw ground beef for the organism in October 1994. Industry challenged the policy in Texas Food Industry Association v. Espy, and the district court refused to block the policy in 1994, with the Fifth Circuit affirming in 1995, on reasoning that many consumers regard ground beef cooked rare or medium as properly cooked, so ordinary cooking does not reliably destroy the organism. In 1999 the policy was extended to non-intact beef products, and in September 2011 six additional Shiga toxin producing strains were announced as adulterants in raw non-intact beef, with routine testing beginning June 4, 2012.

Q: Can USDA order a meat recall?

No. The Food Safety and Inspection Service has no mandatory recall authority for meat, poultry, or egg products. A recall is the establishment voluntary action, often taken at the request of the service, which verifies effectiveness, publicizes the recall, and classifies it as Class I, Class II, or Class III according to health risk. Around that absence, the service holds other tools, including administrative detention of product in commerce for up to 20 days, judicial seizure and condemnation through the United States Attorney, retention of product inside establishments, refusal of entry for imports, and withholding or suspension of inspection. The Food and Drug Administration received mandatory recall authority over most food in section 206 of the FDA Food Safety Modernization Act of 2011. Bills to grant the Agriculture Department a comparable power were introduced repeatedly and were not enacted through February 1, 2021.

Q: Who reviews USDA rules before they are issued?

Food Safety and Inspection Service rules are issued by the Secretary of Agriculture through the service under the Administrative Procedure Act, with proposed rules published in the Federal Register for public comment and final rules responding to comments received. Significant rules undergo review by the Office of Information and Regulatory Affairs in the Office of Management and Budget under Executive Order 12866, issued in 1993. Major rules of the Department of Agriculture also undergo internal review by the Office of the Chief Economist and the Office of Risk Assessment and Cost-Benefit Analysis, created by the Department of Agriculture Reorganization Act of 1994, which examines regulatory impact analyses and risk analyses. Rulemaking documents are posted through Regulations.gov. Review therefore tests both the legal basis of a rule and the quality of the cost, benefit, and risk analysis offered in support of the rule.

Q: What does amenable mean for a meat species under federal inspection law?

Amenable describes a legal status rather than a biological judgment. A species is amenable when one of the inspection statutes subjects that species, and products derived from that species, to mandatory federal inspection by the Food Safety and Inspection Service. Slaughter and processing of an amenable species must occur under inspection as a condition of lawful commerce, with inspectors present during slaughter operating hours. A species outside the statutory definitions is not amenable, however similar that species may be to a covered species in preparation or sale. Exotic and game meats that are not amenable fall to the Food and Drug Administration under the general food statute, although the inspection service offers voluntary inspection on a fee for service basis for some non-amenable species. Congress can change the status of a species by statute, as the farm act of 2008 did for Siluriformes fish, including catfish.

Q: How does federal inspection differ for shell eggs and egg products?

Shell eggs and egg products fall on different sides of the federal line, although both begin as the same food. Shell eggs in production, transport, storage, and retail fall under the Food and Drug Administration. Egg products, meaning eggs removed from the shell and processed into liquid, frozen, or dried form, fall under the Food Safety and Inspection Service under the Egg Products Inspection Act of 1970, with continuous inspection of egg products plants. The dividing principle is the form in which the food is presented and the statute under which Congress placed that form. A third federal function meets the same product without changing safety jurisdiction, because the Agricultural Marketing Service of the Department of Agriculture grades shell eggs for quality only, and grading is not a safety determination. A rule published on October 29, 2020, required egg products plants to adopt Sanitation Standard Operating Procedures and hazard analysis systems on compliance dates in late 2021 and 2022.

Q: Why does catfish fall under a different agency from other seafood?

Catfish fall under the Food Safety and Inspection Service because Congress moved them there by statute. Fish of the order Siluriformes, which includes catfish, were transferred to the inspection service by the farm act of 2008. The transfer was implemented through an inspection service final rule in 2015, with inspection beginning in 2016. All other seafood remained with the Food and Drug Administration under the Federal Food, Drug, and Cosmetic Act. A catfish fillet and a fillet of another fish may be processed, frozen, shipped, and cooked in comparable ways, yet the catfish enters a regime of establishment inspection built originally for livestock and poultry, while the other fillet enters a regime of periodic, risk based facility inspection. The episode shows that the jurisdictional boundary is a legislative artifact. A single statutory transfer can move one defined food category across departments, inspection philosophies, and enforcement systems while neighboring categories stay where earlier statutes placed them.

Q: What happens when an establishment fails a sanitation requirement?

Sanitation failures are directly enforceable by the Food Safety and Inspection Service, unlike failure of a Salmonella performance standard standing alone after the Fifth Circuit decision in Supreme Beef Processors. The Pathogen Reduction and Hazard Analysis and Critical Control Point rule of July 25, 1996, required establishments to maintain Sanitation Standard Operating Procedures, written procedures stating how sanitation would be carried out in daily operation. An establishment that fails to maintain sanitary conditions or to follow those procedures can face agency action grounded in the sanitation requirements themselves, including withholding or suspension of inspection in circumstances the statutes and regulations specify. Suspension is a serious step because an establishment cannot lawfully slaughter without inspection coverage and cannot apply the marks of inspection while inspection is suspended. Salmonella sampling results can guide inspectors toward establishments requiring closer examination, but the violation cited for enforcement rests on the sanitation or hazard system failure rather than on the prevalence count alone.

Q: How long can the inspection service detain product in commerce?

Administrative detention under the Federal Meat Inspection Act is a temporary hold, not a removal. Under 21 U.S.C. 672, an authorized representative of the Secretary of Agriculture may detain meat or a meat food product found in commerce for a period not to exceed 20 days, pending action under the seizure provisions or notification of an appropriate authority, when the representative has reason to believe the product is adulterated or misbranded or was not inspected and passed. Parallel detention provisions apply to poultry and egg products. A detained lot cannot move as ordinary commerce while detention holds, which gives the service time to complete testing, assemble a basis for seizure, or secure voluntary action by the establishment. Detention does not condemn product, does not retrieve product already sold to consumers, and does not reach product the agency cannot locate. The Food and Drug Administration holds a separate administrative detention power for food under section 304(h) of the Federal Food, Drug, and Cosmetic Act.

Q: What is the difference between detention and seizure of adulterated product?

Detention and seizure operate at different points in the enforcement sequence and carry different legal weight. Administrative detention is an agency hold on product in commerce, lasting up to 20 days for meat under 21 U.S.C. 672, with parallel provisions for poultry and eggs, imposed when an authorized representative has reason to believe product is adulterated or misbranded. Detention holds product in place while the Food Safety and Inspection Service decides on a further step. Seizure and condemnation are judicial. When the Secretary seeks permanent control of product, a proceeding against the product itself is brought through the United States Attorney, and a court decides whether the statutory basis for condemnation has been established. Condemnation can end in destruction or in another disposition ordered by the court, and the route does not depend on establishment consent. Seizure is therefore a genuine compulsory power, but a lot by lot judicial power, slower and narrower than the distribution wide removal a cooperative establishment can carry out through a recall.

Q: How do state inspection programs operate beside the federal program?

State inspection programs operate inside the federal framework rather than in place of the federal standard. Under the inspection statutes, a state may run a program described in the statutory phrase as at least equal to the federal program, enforcing requirements equal to or stronger than federal requirements under cooperative arrangements with the Department of Agriculture. Such programs operate in roughly half the states and cover product handled within the state maintaining the program. Product inspected under a state program is, in the general structure of the system, confined to commerce inside that state. The Cooperative Interstate Shipment program, established by the farm act of 2008, created a limited exception, permitting selected establishments inspected under state programs to ship product across state lines under federal coordination. The exception is selective by design, because interstate movement was the historical justification for the federal inspection presence, and every departure from intrastate confinement of state inspected product has required a deliberate statutory decision by Congress.

Q: Why must a foreign inspection system be equivalent before imports enter?

Equivalence is the border form of the continuous inspection model. The Food Safety and Inspection Service does not admit meat, poultry, or egg products on the strength of inspection applied shipment by shipment at a port alone. A foreign inspection system must first be determined equivalent to the United States system, with laws, controls, and inspection performance that achieve the same level of protection, before product from that system becomes eligible for entry. Imported product is then reinspected at import establishments inside the United States, and product that fails reinspection or comes from an ineligible system or establishment can be refused entry. The exporting country supplies the continuous presence under its own inspectors, and the United States accepts the resulting product because the foreign system is held to deliver protections equivalent to those delivered by federal attendance in a domestic plant. The Food and Drug Administration applies a contrasting border regime built on prior notice and, after the modernization act of 2011, foreign supplier verification by importers.

Q: What role do hazard analysis plans play after a plant is inspected?

Hazard Analysis and Critical Control Point systems, made mandatory by the final rule of July 25, 1996, assign the analytical work of food safety to the establishment that produces the food. Each establishment must analyze its operation for hazards reasonably likely to occur, identify the points where those hazards can be controlled, operate controls at those points, monitor performance, and correct deviations. The Food Safety and Inspection Service verifies the system rather than performing the system, examining whether the plan addresses the hazards of the operation, whether controls are carried out, and whether records show the system functioning. Phase in ran by establishment size, with large establishments covered from January 26, 1998, small establishments from January 25, 1999, and very small establishments from January 25, 2000. The structure proved adaptable after adoption, because in 2002, after a large ground beef recall, the service directed ground beef establishments to reassess their plans for Escherichia coli O157:H7 rather than rewriting a single government procedure for the whole sector.

Q: How are meat and poultry labels approved before products reach stores?

Labels for meat, poultry, and egg products require prior approval by the Food Safety and Inspection Service before covered product reaches the market, subject to generic approval expanded over time for labels that fit established categories. Prior approval gives the service a checkpoint before commerce at which claims, ingredient statements, handling instructions, and other label elements pass agency review, and a label that has not received a required approval cannot lawfully be used. The arrangement differs from food labeling under the Food and Drug Administration, which operates largely after placement in commerce, so federal correction of an ordinary food label typically follows sale rather than preceding sale through a comparable approval gate. Label jurisdiction follows product jurisdiction, so a recipe change that moves a mixed product across the small meat or poultry thresholds in agency practice can also move the label from the post market system into the prior approval system.

Q: Why has Congress not created a single food safety agency?

Proposals for a single food safety agency have been made and refused over a long record. The Government Accountability Office has recommended a single, independent food safety agency in various forms since the late 1970s, placed federal food safety oversight on the High-Risk List in 2007, and described oversight as fragmented across 15 federal agencies administering more than 30 statutes. The Safe Food Act, introduced repeatedly from 2005 onward by Senator Durbin and Representative DeLauro, would have created a single Food Safety Administration, and a presidential budget proposal in 2015 sought a single agency inside the Department of Health and Human Services. None was enacted through February 1, 2021. Opposition rested on committee jurisdiction, because agriculture committees guard Department of Agriculture programs, on statutory incompatibility, because mandatory continuous inspection does not merge cleanly with a periodic facility regime, on disruption to thousands of inspectors whose presence establishments require in order to operate, and on the argument that coordination failures are management problems a merger does not by itself repair.

Q: What can inspectors do when a plant operates under insanitary conditions?

Inspectors act first inside the establishment, where federal control is strongest. Product that appears adulterated or otherwise ineligible to move can be retained within the plant under agency control while status is resolved, and product that cannot lawfully be brought into compliance can be condemned or destroyed under applicable procedures rather than shipped. Insanitary conditions also support withholding or suspension of inspection, and suspension without a prior hearing is available for certain egregious conditions, subject to notice and procedures giving the establishment an opportunity to respond. Suspension stops lawful slaughter and prevents use of the marks of inspection, so the ordinary business of the plant halts. The statutory language behind such action includes 21 U.S.C. 601(m)(4), under which product prepared, packed, or held under insanitary conditions whereby it may have been rendered injurious to health is adulterated, although the Fifth Circuit in Supreme Beef Processors read the word rendered to require a deleterious change occurring inside the establishment, which limited the use of Salmonella prevalence in incoming raw material as proof of such a condition.