Companies test on animals by exposing them to chemicals, drugs, or product ingredients and then monitoring for toxic reactions, organ damage, or death. The specific tests range from single high-dose exposures that last a few hours to repeated daily dosing studies that stretch over 90 days or longer. Mice and rats make up roughly 95% of all laboratory animals, but rabbits, guinea pigs, dogs, and primates are also used depending on the product and what regulators require.
Acute Toxicity Testing
The most fundamental animal test measures what happens when an animal receives a single, relatively high dose of a substance. According to FDA guidelines, each animal is given one dose, observed for one to two weeks for signs of illness or organ damage, and then killed for examination. Researchers look at which organs were affected, how severe the damage was, and whether the animal showed any recovery before the observation period ended.
The classic version of this is the LD50 test, which stands for “lethal dose 50%.” The goal is to find the exact dose that kills half the animals in a test group. Animals receive escalating amounts of a substance until researchers can pinpoint that midpoint. This test has been widely criticized, and modern guidelines from the FDA now emphasize observing symptoms and recovery rather than simply calculating a lethal dose. A quicker version called the “limit test” skips the escalation: researchers give a group of about five animals a set high dose (5 grams per kilogram of body weight), watch them for 14 days, and if none die, no further testing at higher doses is needed.
Repeated Dose Studies
For chemicals, pesticides, and food additives that people might encounter over long periods, regulators require studies where animals receive a substance daily for weeks or months. The standard protocol from the Organisation for Economic Co-operation and Development (OECD) calls for a 90-day oral toxicity study in rodents, typically rats. At least 20 animals per test group (10 males and 10 females) receive the substance daily at three or more concentration levels, either mixed into food or water or delivered directly into the stomach through a tube.
Throughout the 90 days, researchers weigh the animals at least weekly, track food and water intake, run blood tests, perform eye exams, and collect urine samples. At the end, every animal undergoes a full necropsy with tissue examination under a microscope. The point is to find a “no-effect level,” the highest dose at which no harmful changes appear. That number then informs the safety limits set for human exposure. A shorter 28-day version often comes first to flag obvious problems before committing to the full three-month study.
Skin and Eye Irritation Tests
Cosmetic and chemical ingredients have historically been tested by applying them directly to animal skin or eyes. The Draize test, developed in the 1940s, involves placing a substance into a rabbit’s eye or onto a patch of shaved skin and scoring the resulting redness, swelling, or tissue damage over several days. Rabbits are used because their eyes produce fewer tears than human eyes, making reactions easier to observe, and because their skin responds in ways considered roughly comparable to human skin.
Skin sensitization tests check whether a substance triggers allergic reactions. Guinea pigs have traditionally been the primary species for these studies. The animal is exposed to a substance repeatedly over weeks, and researchers watch for immune-mediated skin reactions that would suggest the ingredient could cause contact allergies in people.
Pharmaceutical Drug Testing
Before any new drug can be tested in humans, the FDA requires preclinical animal studies to assess its safety. These typically include repeat-dose toxicity studies lasting one to six months, along with evaluations of how the drug moves through the body and whether it affects major organ systems like the heart and liver. For biological drugs such as antibodies, researchers also monitor whether the animal’s immune system mounts a response against the drug itself, which can complicate results.
Pharmaceutical testing often requires two species: one rodent (usually rats) and one non-rodent (often dogs or primates). The logic is that different species metabolize drugs differently, so testing in two gives a broader picture of potential risks. In fiscal year 2024, USDA records show about 40,500 dogs and 60,400 nonhuman primates were held in U.S. research facilities, with the majority used in pain-minimized or pain-free procedures.
Which Animals Are Used and How Many
Mice are by far the most common laboratory animal, making up about 61% of all research animals in large tracking studies. Rats account for another 14%. Together, mice and rats represent roughly 95% of all lab animals. In the United States, however, mice and rats are not covered by the Animal Welfare Act, so they don’t appear in USDA tracking numbers. The species that are tracked paint a detailed picture of what else is involved.
USDA data for fiscal year 2024 recorded about 650,000 animals covered by federal reporting. The largest groups were guinea pigs (around 114,000), rabbits (around 106,000), other mammals (around 115,000), hamsters (around 69,000), birds (around 61,000), and nonhuman primates (around 60,000). Of all tracked animals, about 55,500 were used in procedures where pain was not minimized, meaning they experienced unrelieved discomfort. Hamsters and guinea pigs made up the largest share of that category.
How Welfare Rules Work in Practice
The Animal Welfare Act is the primary U.S. law governing laboratory animals. It requires minimum standards for housing, feeding, veterinary care, and enrichment for covered species. Each research facility must have an Institutional Animal Care and Use Committee that reviews and approves protocols before any experiment begins. The committee evaluates whether the proposed number of animals is justified, whether pain management is adequate, and whether alternatives to animal use were considered.
The law’s biggest gap is that it excludes mice, rats, and birds bred for research, which make up the vast majority of animals used. These species may still receive institutional protections, but they aren’t subject to USDA inspections or reporting requirements.
Cosmetic Testing and Global Bans
Cosmetic animal testing has become increasingly restricted worldwide. Forty-five countries now prohibit it, including all European Union member states, the United Kingdom, Australia, Brazil, Canada, India, South Korea, and Mexico. These bans typically cover both testing finished cosmetic products on animals and testing individual cosmetic ingredients on animals.
In markets where bans are in place, companies rely on existing safety data for well-known ingredients, computer modeling, and cell-based lab tests. However, some countries, notably China, have historically required animal testing for imported cosmetics, creating a tension for brands that want to sell globally while claiming cruelty-free status.
What “Cruelty-Free” Certification Requires
The Leaping Bunny certification, one of the most recognized cruelty-free standards, goes beyond a company’s own practices. To qualify, a company must set a fixed cutoff date after which no animal testing occurred anywhere in its supply chain. Every third-party manufacturer must sign a declaration confirming they did not conduct or commission animal testing on any ingredient, formulation, or finished product. Raw material suppliers must provide their own signed declarations as well.
Companies are also subject to independent audits. Those with gross annual sales of $10 million or more must commission an audit through an accredited firm. Smaller companies must agree to an audit if requested. This supply chain verification is what distinguishes Leaping Bunny from vaguer “not tested on animals” claims, which may only refer to the finished product while ignoring ingredient-level testing.
Alternatives Gaining Regulatory Ground
Organ-on-a-chip technology is emerging as a serious replacement for certain animal tests. These are small devices containing living human cells arranged to mimic the function of a specific organ. Liver chips, for instance, can detect drug-induced liver injury using human, dog, and rat cells on the same platform, allowing researchers to compare species-specific responses without using live animals. Kidney chips have been used to measure drug toxicity in human kidney cells with results closer to what happens inside the body than traditional cell cultures achieve. Heart chips can screen for cardiac side effects by measuring how human heart cells respond to known dangerous drugs versus safe ones.
The FDA has begun outlining how these technologies could formally replace animal studies. One path forward would allow companies to submit a single-species animal study paired with chip-based or computer-modeled data, rather than requiring studies in two species. For some products, the agency has signaled that an entirely non-animal approach could be sufficient if it addresses the same safety questions that animal studies were designed to answer. Computer simulations can now model how a drug distributes through the human body and predict effective dose ranges, potentially replacing some of the animal pharmacology studies that have been standard for decades.