How to tell a sealed jar from a jar with a lid on it

The two questions people run together

Ask whether a jar is safe and you are actually asking two things at once.

Did the jar close? A seal test answers this. It is a mechanical question about a vacuum, a piece of metal and a rubberised gasket, and you can answer it from outside the glass in about two seconds.

Was the food heated enough? No test in your kitchen answers this. It was decided hours ago by the method you used, the recipe you followed, the size of the jar, the density of the pack and your altitude. By the time the jar is cool, that question is closed and the answer is already fixed — you just cannot read it off the lid.

The seal test is easy and the process question is not, so the seal test gets asked to do both jobs. It cannot. Keeping them apart is the single most useful habit on this side of the site, and it is why the last of the source quotes below is on this page at all.

Run the test on a cool jar with the band off

Both halves of that sentence matter.

Cool means the 12 to 24 hours the guide asks for [F-11.2]. The vacuum forms as the contents cool and contract. Prodding a warm lid gives you a reading of a process that has not finished.

Band off matters for two reasons. A band is a clamp, and it is also the thing you must not turn: the guide's instruction not to retighten a lid after processing appears twice, and the reason it gives is that retightening a hot lid can cut through the gasket and cause the seal to fail [F-8.8].

And a clamp, because a band left on and snug can hold a lid flat against the rim whether or not there is any vacuum underneath, and your finger will find a firm lid either way. The guide's own test sequence starts with removing the screw bands for exactly this reason [F-11.1].

Then use one of the three. If you want a single default: look across the lid at eye level. A properly sealed lid is visibly concave, dished down in the centre, and once you have seen a dozen of them you can read a shelf of jars in one pass.

What a failed test means, and what it does not

A jar that fails the test on day one is not a disaster and not a mystery. Some proportion of lids do not take. Check the rim for a nick, which is the usual culprit and is easy to feel with a fingertip.

You then have three tested routes, all from the guide: reprocess within 24 hours with a new lid and the same processing time, freeze it with the headspace adjusted, or refrigerate that jar and eat it soon [F-9.9]. Pick whichever suits the evening. None of them is second best.

What a failed seal does not mean is that the food is dangerous right now. A jar that never sealed has been sitting in air, at room temperature, which is the condition under which spoilage announces itself in the ordinary ways — and the guide's own instruction still applies: do not taste food from a jar with an unsealed lid [F-11.7]. Refrigerate it or throw it away. Do not sample it to decide.

The failure that is worth watching for later

A seal that formed can fail afterwards, and the guide tells you what that looks like: spoilage organisms produce gas, the gas pressurises the food, the lid swells and the seal breaks [F-8.9].

That is a visible event, and it is the reason the bands come off for storage. A swollen lid on a bandless jar is obvious from a metre away. A swollen lid under a screwed-down band may look like nothing at all, because the band is doing what a band does and holding it down.

This is the rare case where the guidance costs you nothing to follow. Take the bands off, wash them, keep them. You get a free spoilage indicator on every jar for the price of a step you were arguably supposed to do anyway.

The honest summary

Three tests, run on a cool jar with the band off, tell you a vacuum is present [F-11.1]. That is a real and useful thing to know, and it is worth being exact about how little it is. It tells you the closure worked. It tells you nothing about the heat the food received. A jar has to have both before it belongs on a shelf, and only one of the two can be checked afterwards — the other was settled by the process you ran, before the lid went on. A jar that fails the seal test after a tested process is a different question again, and it has three tested answers [F-9.9].

So the seal is not a verdict on the food. Jars sealed by methods that never heated the food at all pass every one of those tests [F-10.2]. If you need confidence about the process rather than the closure, it comes from having used a tested method for that specific food — which is a decision made before the lid went on, not a reading taken after.