Milk spoilage after opening

Why does milk keep for weeks unopened but sour within days once the carton is opened?

The carton sat in the fridge for two weeks, sealed, and was fine. You opened it, put it straight back in the same fridge at the same temperature, and five days later it has turned. Nothing about the storage changed. The only event was the removal of a plastic cap.

That should bother us. If cold is what preserves milk, the cold was constant throughout. So the cold cannot be the whole story — and whatever the cap was doing, it was doing something the refrigerator was not.

Start with what pasteurisation actually does. The usual treatment holds milk at about seventy-two degrees for around fifteen seconds — enough to destroy pathogens reliably and knock down the general microbial load, and emphatically not sterilisation. Some organisms survive it, notably heat-resistant spore-formers, and more are picked up afterwards during filling from equipment and air. A sealed carton therefore leaves the dairy with a small live population in it.

Small, but not zero. And here is the step that carries the argument: bacterial growth is exponential, so a population does not creep toward spoilage at a steady pace. It doubles. Spoilage is not a clock running down; it is a population crossing a threshold, the point at which enough acid and enough enzyme-broken protein and fat have accumulated for you to taste and smell it.

Two things then determine how long that takes: how many organisms you begin with, and how fast they double.

Refrigeration attacks the second. Cold slows metabolism enormously — but it does not stop it, and the organisms that matter in milk are exactly the ones least troubled by it. They are called psychrotrophs: cold-tolerant bacteria, Pseudomonas prominent among them, that grow perfectly well at four degrees, just slowly. So the sealed carton is a race between a small starting population and a slow doubling time, and it takes a couple of weeks for the count to climb high enough to notice.

Now open it. Three things happen at once. Room air, the pouring lip, a splashed spoon, the inside of the fridge door all deposit new organisms, raising the starting number sharply. Oxygen enters, and the most aggressive spoilage bacteria in milk are aerobes that grow far better with it. And the carton now gets taken out, left on the counter through breakfast and put back — every excursion a spell at a temperature where doubling is several times faster.

Consider what a higher starting number does to an exponential process. Every tenfold increase in the initial count removes a little over three doublings from the milk's remaining life. Multiply the count by a thousand on opening — easily done — and you have deleted roughly ten doublings, which at fridge temperature is a matter of days. You have not made the milk spoil faster. You have started it much further along the same curve.

Is there a clean way to test this account? There is, and it is sitting in the same supermarket. UHT milk is heated to around 135 degrees for a few seconds and filled into sterile packaging under aseptic conditions. It is commercially sterile: no viable organisms inside, and no route for any to get in. It sits on an unrefrigerated shelf for months. Open that carton and it must go in the fridge and will spoil in a few days, exactly like the pasteurised one.

Notice what that comparison isolates. Same liquid, same fridge, same bacteria available in the kitchen air. The only variable is whether organisms are present and whether they can get in. Take that away and the milk lasts months without cold; restore it and no amount of cold buys more than a few days. Preservation was never chiefly the refrigerator. It was the barrier.

Three refinements are worth stating.

First, "spoilage" is not one process. Souring is lactic acid bacteria fermenting lactose. But the bitter, off, sometimes putrid character of refrigerated milk gone wrong is usually the work of psychrotrophic bacteria breaking down protein and fat with enzymes — a different chemistry with a different taste, and the more common failure mode in modern cold chains.

Second, and this is a genuinely counter-intuitive detail: some of those bacterial enzymes are more heat-stable than the bacteria that made them. Raw milk held too long before processing can therefore accumulate enzymes that survive even UHT treatment and slowly gel or embitter the product months later. What limits UHT shelf life is often chemistry, not microbes.

Third, the printed date is a quality estimate under assumed handling, not a property of the liquid. It presumes an unbroken chain at or below about four degrees; a fridge running at eight is a different experiment from the one the date was based on.

None of this makes refrigeration unimportant — it is what turns a few hours into a couple of weeks. The point is only that it slows the process without arresting it, exactly as you would expect of something acting on a growth rate rather than on the population itself.

• Why raw milk sours pleasantly while pasteurised milk more often turns bitter and putrid.
• How cold-chain temperature logging changed the way dairies set and defend their printed dates.

Every term the collection defines is gathered in the glossary.