Cheese ripening

Why does deliberately letting microbes grow in milk make it keep for years when fresh milk sours in days?

Leave milk on the counter and in a day or two it is undrinkable. Add bacteria to that same milk on purpose, cut it, press it, and put it in a cellar, and two years later it is not merely edible but better than it was.

The obvious reading — that cheese is milk that spoiled in a way we happen to enjoy — cannot be right, because spoilage does not stop. Sour milk goes on getting worse. Cheese arrives somewhere and stays. What is different about the destination?

Ask what a spoilage organism needs. Water it can use, a fermentable sugar or protein it can break down, a pH it tolerates, a workable temperature, and somewhere to grow before a competitor arrives. Remove any one and growth stops, whatever else is present. Now walk the making of a hard cheese and watch each item disappear.

The starter culture goes in first, in enormous numbers, and it converts lactose to lactic acid. Milk is about 4.8 per cent lactose (recalled) and sits near pH 6.7. The acid drops the pH toward 5 (recalled for a cheddar-type cheese), which is where the casein micelles lose their charge, stop repelling each other, and coagulate. So one action does two jobs, acidification and curd formation, and only one of them is preserving anything.

Then the curd is cut and the whey drained. Milk is around 87 per cent water; a hard cheese ends near 35 per cent moisture (recalled). But the water leaving takes the dissolved things with it, and lactose is dissolved — so most of the fermentable sugar goes out in the whey, and the starter finishes the rest within days. A mature hard cheese has essentially no lactose left, a fact you can check from the other end, since aged hard cheeses are tolerated by people who cannot drink milk.

Then salt. Salt does not have to poison anything; it lowers water activity, the availability of water rather than its amount. Cheesemakers control it not as a share of the cheese but as a share of its moisture — salt-in-moisture, around four to five per cent in cheddar (recalled) — because it is the concentration in the water phase that microbes experience.

Now count what a spoilage organism arriving on day thirty finds: no fermentable sugar, a pH near 5, water partly unavailable, an interior with almost no oxygen, and every surface already occupied. Not a hostile environment in the sense of a poison. An exhausted one.

So what is ripening, if growth has stopped? Mostly not growth at all. Enzymes — from the coagulant, from the milk, and from starter cells as they die and lyse — slowly break proteins into peptides and amino acids and split some fats into free fatty acids. That is where flavour and texture come from. And here is the elegant part: it is slow because the environment is hostile. The same conditions that keep spoilers out slow the chemistry to the pace of months, which is why a cheese matures and keeps at once. They are one process seen from two sides.

Three folk accounts need ruling out, each for a reason. "The good bacteria kill the bad ones" attributes the work to antimicrobial warfare. Some starters do make bacteriocins — nisin is the famous one — but they are not the main mechanism, and the test is simple: cultures producing none preserve a cheese perfectly well. "Salt kills the microbes" mistakes the lever; the concentrations used are far below lethal for many organisms, and the effect is on water availability — which is why a dried food with nothing added keeps for the same reason. And "cheese is controlled spoilage" gets the shape wrong: spoilage is an open succession in which each population prepares the ground for a worse one, whereas cheesemaking is a succession driven to a dead end where nothing following has anything to consume.

The falsification test the dairy case provides for free. My account says preservation comes from removing growth conditions, not from the presence of microbes. So a cheese from the same milk and culture, but not driven far on moisture, salt and acid, should spoil like milk — and fresh mozzarella and ricotta, high-moisture, lightly salted and barely acidified, keep for days rather than years. Microbes held constant, conditions varied, shelf life tracking the conditions. A second test from industry: when a bacteriophage attacks the starter and acidification fails, the vat does not merely make a cheese that tastes flat, it makes one that goes off, because the pH never fell and the lactose was never consumed. If the microbes were themselves the preservative, that would be a flavour problem, not a safety one.

Where this sits next to its neighbours: food-preservation.md derives the general lever — salting, drying and sugaring all lower water availability rather than water content — and cheese is an instance of it, differing at one fixed point: cheese adds a step drying does not have. A raisin is preserved passively, by taking something away; a cheese first recruits a living population to consume the fuel and drop the pH, and only then removes the water. milk-spoilage-after-opening.md starts from the same litre and reaches the opposite outcome, and the pivot there is starting count against doubling time — an opened carton is a rich, neutral, wet medium in which a small population grows exponentially to a threshold. Cheesemaking never lets that competition begin, using an overwhelming starter added first that strips the medium bare.

The load-bearing claim is that preservation comes from removing growth conditions — fermentable substrate, pH, water activity — rather than from antagonism between organisms. If bacteriocin-free starters failed to preserve, or a cheese with lactose still in it kept as well as one without, this account would be wrong.