People are told to cook chicken to 74 °C / 165 °F and are rarely told why. The number is not a threshold where bacteria suddenly die. It is the point where they die *instantly* — and understanding that changes how you read every other number on this site.
Pathogens are killed by time at temperature, not by a temperature
Heat kills Salmonella progressively. Hold it at a given temperature and its population falls at a predictable rate, so the same reduction can be reached either quickly at a high temperature or slowly at a lower one. Food-safety targets are therefore written as a *log reduction* — a 7-log reduction means the population is cut to one ten-millionth of what it was.
US regulations set that bar differently by product: a 7-log reduction of Salmonella for poultry, 6.5-log for roast and corned beef, and 5-log for cooked patties (9 CFR 381.150, 318.17 and 318.23 respectively).
74 °C / 165 °F is simply where the required poultry reduction happens with no holding time at all. In the USDA's own words, "A 7-Log reduction of Salmonella is achieved instantly at internal temperatures in which the holding time is 0 seconds." For meat the equivalent threshold is 70 °C / 158 °F: "The required Log reductions are achieved instantly (0 seconds) when the internal temperature of a cooked meat product reaches 158°F or above."
So is chicken at 68 °C safe?
In principle, yes — if it is genuinely held there long enough. That is what the USDA's time-temperature tables quantify: for every internal temperature below the instant threshold, a dwell time that achieves the same reduction.
In practice the qualifier does almost all the work, and this is where most writing on the subject goes wrong. Those tables are not consumer advice. They are part of a guideline that states it "applies to small and very small meat and poultry official establishments", and they are only valid under conditions a domestic kitchen does not reproduce.
What the tables assume that your kitchen does not provide
- Monitored humidity. An establishment using the tables "must address humidity", monitoring it with wet and dry bulb thermometers or a humidity sensor — recommended for every batch. Dry surfaces let bacteria survive heat that would otherwise kill them.
- A controlled come-up-time. The product's journey through the danger zone is itself a monitored parameter, capped at six hours between 10 °C and 54 °C / 50 °F and 130 °F.
- Known fat content. Fat protects bacteria from heat. The USDA publishes separate columns for fat levels from 1% to 12%, citing research showing higher fat raises Salmonella's heat resistance. Where the fat content is unknown, the guidance is to use the 12% figures — the most conservative.
- Known species. Chicken and turkey get separate tables, because heat tolerance differs between them.
- A temperature you can actually hold. A dwell time is meaningless if the temperature drifts. An oven cycling around a setpoint does not hold a piece of meat at 68 °C for a measured number of minutes.
The exception worth knowing: immersion and sealed bags
The humidity requirement exists because a drying surface protects bacteria. The USDA accordingly exempts processes where drying cannot happen — among them full immersion in liquid and cooking in a sealed, moisture-impermeable bag.
That is a precise description of sous vide. A water bath also solves the other problem, because it genuinely holds a temperature rather than cycling around one. Of all domestic cooking methods, sous vide is the one where time-and-temperature reasoning transfers most honestly — which is why sous-vide cooks talk about pasteurization while roasting cooks talk about a single number.
This is not a licence to improvise. It means that if you want to work below the instant threshold, the equipment has to do what the tables assume, and you should be reading the source tables rather than a number half-remembered from a forum.
Why this site publishes single numbers
Every poultry page here gives one safe minimum, because that is the number that is correct without a stopwatch, a humidity reading and a known fat percentage. It is the honest default for cooking in an oven or a pan.
It is also why the doneness scales on beef and lamb pages go *below* the USDA safe minimum without contradiction: an intact cut carries its bacteria on the outside, where searing destroys them, so the interior temperature answers a question about texture rather than safety. Ground beef gets no scale at all, because grinding moves that surface contamination throughout. The methodology page covers that distinction in full.
Where a national authority sets a different figure from the USDA, this site follows the local authority and says so on the page.
Sources, and a caveat about them
The quotations above come from the USDA FSIS Cooking Guideline for Meat and Poultry Products (Revised Appendix A), dated 14 December 2021, which replaced all previous versions.
This page deliberately does not reproduce the numeric tables. They are distributed as images rather than text, so any transcription circulating online — including earlier versions still widely quoted — cannot be checked easily against the current document, and several superseded editions remain in circulation. For a figure you intend to rely on, read it from the source above rather than from anyone's retyping of it, this site included.
The guideline itself notes that its contents "do not have the force and effect of law". It describes how processors demonstrate compliance; it is not a set of instructions for dinner.