🥛 Milk Quality and Testing

Before you can test milk you have to know what is supposed to be in it. Milk is not one substance. It is water with several different things dissolved and suspended in it, and every test in this module works by measuring one of those things or by measuring what bacteria have done to them.

Here is the average composition of whole cow milk. This is a United States reference figure, given here for the general chemistry of milk rather than as an African standard:

• Water: 87.7 percent
• Lactose (milk sugar): 4.9 percent
• Fat: 3.4 percent
• Protein: 3.3 percent
• Minerals, also called ash: 0.7 percent

An FAO dairy composition reference gives a broadly consistent general range: fat about 3 to 4 percent, protein about 3.5 percent, and lactose about 5 percent. The same FAO reference notes that water content across all dairy species runs from about 83 percent in yak milk to about 91 percent in donkey milk, with cow milk sitting inside that band.

So roughly seven-eighths of what you carry to the collection point is water. That single fact explains a great deal - why milk is heavy and bulky relative to its value, why transport is such a large share of a smallholder's costs, and why processing into a concentrated product such as ghee or cheese is commercially attractive.

Everything that is not water is what the buyer is actually paying for, and the components behave differently.

Fat is the component that varies most, both between cows and within a single milking. It is suspended in the milk as globules rather than dissolved, which is why cream rises. It is also the component most dairy pricing systems reward, because fat drives the yield of butter, ghee and full-fat products.

Protein is the component that sets curd and makes cheese, and it is the component the alcohol test in Lesson 3 is really examining.

Lactose is the sugar. It is the component bacteria eat, turning it into lactic acid, which is why old milk is sour milk and why acidity is a proxy for age. It is also what makes fermented products possible.

Minerals matter mainly because they affect how the protein behaves under heat, and because milk is a genuinely good source of calcium in a household diet.

Now the commercially important part, and it is one of the most useful sourced facts in this whole course.

Gross milk composition varies by breed. Specifically, Bos indicus cattle - the zebu-type cattle that make up most indigenous African herds - typically have higher milkfat content than Bos taurus, the European taurine breeds. Zebu milk fat is reported as reaching as much as 5.5 percent.

Read that against the average of about 3.4 percent above, and against everything you know about yield. The zebu cow gives fewer litres per day. That is true and it is not in dispute. But her milk is often richer in fat than the milk of the exotic cow she is unfavourably compared with.

This is direct ammunition against the common assumption that indigenous milk is simply worse milk. It is not worse. There is less of it, and it is richer. Whether that combination is good or bad for you depends entirely on one thing: how your buyer pays.

If your buyer pays a flat price per litre, composition earns you nothing, and only volume counts. If your buyer pays on fat or on solids content, then the zebu or zebu-cross cow is being paid for her actual advantage, and your management decisions should change accordingly.

This matters for processing too. If you make ghee, fat content directly drives your yield, because ghee is clarified butterfat and nothing else. A litre of 5.5 percent milk contains meaningfully more of the product you are trying to make than a litre of 3.4 percent milk. The same logic applies to butter and to full-fat cheeses.

So the practical task from this lesson is to go and find out how you are paid. Ask your buyer directly: do you pay per litre regardless of quality, or do you pay on fat, on solids-not-fat, or on any composition measure? Write the answer down. It determines which of the improvements in this whole course will actually put money in your pocket, and it is astonishing how many farmers sell milk for years without ever asking.

The lactometer is the first tool on any milk collection bench, and it exists to answer one question: has water been added to this milk?

How it works is straightforward physics. A lactometer measures specific gravity, which is another way of saying density - how heavy the milk is for its volume, compared with water. Milk is denser than water, because of the dissolved lactose, protein and minerals it carries. Add water to milk and you dilute all of those, and the density falls. The lactometer floats in the sample and how deep it sinks tells you the density.

Here are the figures.

The normal specific gravity range for whole milk is 1.028 to 1.034, measured at 15 degrees Celsius. In practice, at a collection point, people record only the last two digits, so you will hear readings called simply "28" to "34". A reading in that band is normal. A reading below it is a flag.

Now the part that is skipped more often than any other step in milk testing, and which invalidates the whole test when it is skipped: temperature correction.

Density depends on temperature. Warm milk is less dense than cold milk of exactly the same composition, so if you take a reading on warm milk and compare it against a table written for 15 degrees, you will read low and accuse an honest farmer of watering. The sourced correction table gives adjustments ranging from minus 0.007 at 17 degrees Celsius to plus 0.011 at 24 degrees Celsius, relative to the 15 degree reference.

The practical rule is absolute: a lactometer reading must never be taken or interpreted without also recording the milk's temperature. If you do not have a thermometer, you do not have a lactometer test - you have a number that means nothing. Buy the thermometer with the lactometer, always. They are one instrument in two pieces.

How to run the test:

1. Mix the milk gently but thoroughly first. Milk that has been standing has cream at the top and thinner milk below, and a sample from either will read wrong.
2. Pour the sample into a tall cylinder, filling it enough that the lactometer floats free without touching the bottom or the sides.
3. Lower the lactometer in gently and let it come to rest. Do not push it down and let it bounce.
4. Read the scale at the milk surface, at eye level.
5. Take the milk temperature with a thermometer at the same time and write both numbers down together, never one without the other.

Now interpreting it, and this is where honest teaching matters.

A reading below the normal range is consistent with added water. That is the reason the test exists. But it is not proof, and the reason is genuinely confusing until you think it through.

Fat is lighter than the watery part of milk. So removing fat - skimming - actually raises the density of what is left. Adding water lowers it. That means the two commonest forms of cheating push the reading in opposite directions, and a milk that has been both skimmed and watered can come back reading perfectly normal, because the two effects cancel out.

So a lactometer reading alone cannot distinguish watered milk from skimmed milk, and cannot detect a sample that has been both. This is exactly why platform testing uses several tests together rather than any one test alone, and why a collection point that owns only a lactometer is less protected than it believes.

One more honest limit. You will hear people quote a formula converting a lactometer reading directly into a percentage of added water. A precise sourced figure for how many points the reading drops per percent of water added was not available for this course, and this module will not print one. Do not teach or use a specific conversion formula as fact on the strength of hearsay. Use the lactometer for what it is sound for: is this reading inside the normal band at the recorded temperature, or is it not?

And finally, remember what a low reading might honestly mean. Milk genuinely low in fat and solids - from a cow late in lactation, a cow in poor condition, or a cow that is ill - can read low without anybody having added anything. Rinse water left in a can from washing does the same thing accidentally. A low reading is the start of a conversation, not the end of one.

The alcohol test is the fastest freshness and stability check on the bench. It takes seconds, it needs no heat and no water bath, and it answers a question that matters commercially: will this milk survive being heated?

That is what it is really testing. Alcohol destabilises milk protein. Milk whose protein is already close to the edge of stability - because it has gone acid, or because its mineral balance is abnormal - will coagulate when alcohol is added, producing visible flakes or clots against the side of the tube. Stable, fresh milk stays smooth.

Why does that matter? Because milk that coagulates with alcohol will also tend to coagulate when heated, and heating is what a processor does. Milk that curdles in the pasteuriser is a plant problem, not just a quality problem, which is why processors and collection points care about this test so much.

The procedure:

1. Mix equal volumes of milk and the alcohol solution in a test tube - for example 2 millilitres of each.
2. Agitate gently. Do not shake violently.
3. Look at the tube, and at the film left on the glass as you tilt it, for coagulation - flakes, clots or grains.

The alcohol concentration:

A 68 percent ethanol solution by weight is the standard - for example, 68 millilitres of 96 percent alcohol mixed with 28 millilitres of distilled water. Alternatively, 75 percent ethyl alcohol by volume is used, depending on the local protocol.

That "depending on the local protocol" is not vagueness, it is important. Different collection points and different countries use different concentrations, and a stronger alcohol is a harsher test that will fail milk a weaker one passes. So if your milk fails an alcohol test at a collection point, one entirely legitimate question is what concentration they are using. Ask, and write it down, so that you are comparing like with like when you test your own milk at home.

The interpretation:

Milk with an acidity over about 0.21 percent, expressed as lactic acid, fails the test. Visible coagulation means the milk is too acidic - too old, or from an abnormal udder - to be processed safely, and it should be rejected.

That single sentence contains the two different reasons milk fails, and you must not confuse them.

The first reason is age. Bacteria have been eating the lactose and producing lactic acid. The milk is old, or it has been warm, or both. This is a hygiene and cold chain failure, and it is the common case. The whole of the module after this one is about preventing it.

The second reason has nothing to do with age at all. Certain abnormal milks are unstable even when perfectly fresh, and they will fail the alcohol test straight from the cow. The named cases are colostrum, late-lactation milk, and mastitis milk.

Colostrum is a completely different fluid from ordinary milk, far higher in protein and minerals, and it is not stable to alcohol. This is one more reason - beyond everything you learned about the calf's need for it - that colostrum and the first days' milk after calving do not go into the sale can.

Late-lactation milk, from a cow near the end of her lactation, shifts in composition and mineral balance and can fail.

Mastitis milk fails because inflammation changes the salt balance of the milk, and it is worth pausing on this. A cow with clinical mastitis that you missed at the strip cup can fail your whole delivery on an alcohol test. That is one more reason the free strip cup check in the milking module earns its place.

So the correct reading of an alcohol test failure is: this milk is not fit to be processed, and now I must work out why. Was it old or warm - a cold chain problem? Or is one of my cows contributing abnormal milk - an udder health problem? Those two answers lead to completely different actions, and the test itself will not tell you which. Your own records will.

One last practical caution. Alcohol at these concentrations is flammable and is not drinkable. Keep it stoppered, keep it away from any flame or heat source, label the bottle clearly, and keep it out of reach of children. Do not store it in a container that has ever held anything drinkable.

The resazurin test is the closest thing a collection point has to a bacteriological test without a laboratory. It takes ten minutes and it estimates how heavy the bacterial load in the milk is.

The principle is elegant. Resazurin is a blue dye. Living bacteria, as they respire, chemically reduce it, and as they do so it changes colour - through purple, to pink, and eventually to white. The more bacteria there are and the more active they are, the further the colour travels in a fixed period of time. So you add the dye, hold the sample warm for exactly ten minutes so the bacteria are working at a standard rate, and then read where the colour has got to.

You are not counting bacteria. You are watching them work and inferring how many there must be.

The procedure:

1. Add 1 millilitre of resazurin solution to 10 millilitres of milk in a stoppered tube.
2. Invert the tube gently twice, within about 4 seconds, to mix.
3. Place the tube in a water bath at 37 to 38 degrees Celsius for exactly 10 minutes.
4. Read the colour against a reference disc.

Every number in that procedure earns its place, and this is a test that punishes carelessness.

The 1 to 10 ratio must be right, because the amount of dye relative to the milk sets the whole colour scale.

The gentle double inversion, within about four seconds, mixes the dye without whipping air into the sample. Shaking a tube of milk hard changes the oxygen in it and therefore changes how the dye behaves.

The 37 to 38 degree water bath matters because bacterial activity is temperature-dependent. Bacteria at 20 degrees work slowly and the sample will read better than it deserves. Bacteria at 45 degrees work fast and the sample will read worse. The temperature is what makes results from different days comparable.

And exactly 10 minutes means exactly. The colour keeps changing. A tube read at fifteen minutes will look worse than the same tube at ten. Use a clock, not a feeling.

Now the scale. The result is read against a reference disc with seven grades:

• Disc 6, blue: Excellent
• Disc 5, light blue: Very good
• Disc 4, purple: Good
• Disc 3, purple-pink: Fair
• Disc 2, light pink: Poor
• Disc 1, pink: Bad
• Disc 0, white: Very bad

Notice the direction. High disc number and blue is good. Low disc number and pink or white is bad. Milk that has gone white in ten minutes at 37 degrees is carrying an enormous bacterial population.

How to use the result honestly:

This test is a screen, not a diagnosis. It flags milk with a high bacterial load for rejection or downgrading. It does not tell you which organism is present, where it came from, or whether it is dangerous. A poor grade is a signal to go and look at your own process, and there are only really two places to look.

The first is hygiene at milking. Bacteria that were introduced from dirty hands, a shared cloth, a dirty udder, or a residue-filmed container are in the milk from the moment it was drawn, and they had a head start.

The second is the cold chain. Bacteria that were few at milking multiply rapidly in warm milk. A perfectly clean milking followed by three hours standing in the sun produces a poor resazurin grade just as reliably as a dirty milking does.

So if you get a bad grade, the diagnostic question is: was this milk dirty when it was drawn, or did it get warm afterwards? Test one batch handled with obsessive hygiene and cooled fast, against your ordinary batch, and the comparison will tell you which of the two is your real problem. That is worth doing once properly, because the two problems have entirely different fixes and it is expensive to fix the wrong one.

One caution about equipment. If you do not have a proper water bath, a pan of water on a low heat with a thermometer in it can be made to hold 37 to 38 degrees, but it takes attention. What you cannot do is guess the temperature or the time and still call the result a resazurin test. Without the thermometer and the clock, you are looking at a coloured tube, not running a test.

This lesson covers the simplest test of the four and then puts all four together into a working bench.

The clot-on-boiling test is the crudest and fastest check there is, and its great virtue is that it needs no reagent at all. Only heat.

The procedure: heat roughly 5 millilitres of milk in a test tube for up to 4 minutes, then rotate the tube and examine it for precipitated particles or clot formation.

The interpretation: milk that clots on boiling generally has an acidity above about 0.22 percent, expressed as lactic acid, and must be rejected.

That is the whole test. Note that it works on the same principle as the alcohol test - heat, like alcohol, destabilises protein that is already close to the edge - and note that its threshold, about 0.22 percent, sits just above the alcohol test's threshold of about 0.21 percent. That is exactly what you would expect. The alcohol test is slightly more sensitive; it catches milk a little earlier. The clot-on-boiling test catches milk that is further gone.

So the practical relationship between them is this: if you have alcohol, use the alcohol test, because it is more sensitive and faster. If you have nothing but a flame and a tube, use clot-on-boiling, because it is far better than nothing and it will catch genuinely sour milk before it goes into a can with everybody else's. A collection point with no equipment at all can start here, today, for the cost of a tube.

Now the temperature figures a bench works to, because a receiving bench checks temperature as well as chemistry.

Milk received at a chilling station should be below 7 to 8 degrees Celsius. That is the reference receiving figure.

And milk held at 0 to 4 degrees Celsius shows no detectable increase in bacterial count within 24 hours. That is the storage figure, and it is the reason the whole cold chain exists. Cold does not kill the bacteria in milk. It stops them multiplying, and stopping them multiplying is enough to hold quality for a day.

Those two figures are the bridge to the next module, which is entirely about how you actually achieve them.

Now sequence the bench. Four tests, each answering a different question, run in a sensible order because some are fast and some cost reagent.

First, and before any test at all: look and smell. Milk that is visibly dirty, has visible clots, is discoloured, or smells sour or of anything other than milk, is rejected without spending reagent on it. This costs nothing and settles a good proportion of bad deliveries.

Second, temperature. Take it with a thermometer and write it down. You need it for the lactometer anyway, and it tells you immediately whether the cold chain held.

Third, the clot-on-boiling or alcohol test. This is your freshness and stability screen and it takes seconds. Whichever you have, do it early, because a milk that fails here is rejected and nothing further needs to be spent on it.

Fourth, the lactometer, with the temperature you already took. This is your added-water check.

Fifth, where the collection point has the equipment, the resazurin test. It takes ten minutes and a water bath, so it is not run on every can at every delivery in most small operations - it is run as a periodic check, or on suppliers whose milk has been a problem.

And then the part that turns testing into management: write it all down, by supplier, by date. A single failed test is an argument. A record showing that one supplier's milk fails on temperature three mornings out of five is a diagnosis, and it lets a cooperative help that member fix a real problem - probably a long uncooled walk to the collection point - instead of simply rejecting his can and resenting him.

That is the correct spirit for a collection-point bench. These tests exist to protect the pooled milk of everybody who delivers into the same can. Used well, they also tell each member exactly what to fix.

This is the most important lesson in the module, and it is about the limits of everything you have just learned.

Here is the sentence to carry away. None of these tests - lactometer, alcohol, resazurin, clot-on-boiling - detects brucellosis, bovine tuberculosis, veterinary drug residues, aflatoxin, or deliberate chemical adulterants such as urea, starch, detergent or hydrogen peroxide.

Read that list again, slowly, because every item on it is a real hazard and none of them changes the density, the acidity, or the ten-minute dye colour of milk in a way these tests would catch.

A cow with brucellosis can give milk that passes every test on the bench. So can a cow with bovine tuberculosis. So can a cow treated with an antibiotic yesterday. Milk deliberately thickened with starch to disguise watering can pass a lactometer test precisely because that is what the starch was added to do.

So the tests in this module are quality-and-freshness screens. They are not food-safety tests and they are not disease tests. Confusing the two is the single most dangerous mistake in this whole subject, and it is a genuinely common one: "she passed the alcohol test, so the milk is fine." It is not fine. It is fresh. Those are different words and they mean different things.

What they are good for is real and worth having. They catch old milk, warm milk, milk that has been diluted, milk that has been produced or handled dirtily, and milk that will fail in a processor's vat. They protect the pooled milk of everyone delivering into the same can from the worst of what any one member might bring. That is exactly what a first line of defence should do, and it does it cheaply.

What they cannot do has to be covered by separate systems entirely:

• Zoonotic disease - brucellosis and bovine tuberculosis - is controlled by heat treatment of the milk, and by whatever herd testing and certification programme your national veterinary authority operates. That is the whole subject of the next module.
• Drug residues are controlled by observing withholding periods, and are screened for, where they are screened for at all, by specialised residue tests that a buyer runs, not by anything on this bench.
• Deliberate chemical adulterants require specific tests for each substance, which are more specialised than anything covered here.

A milk collection point's testing regime should therefore treat platform tests as the routine first line, and treat disease surveillance and drug-residue screening as separate systems running alongside them, not as things the bench covers.

Now the standard itself, and this is where you must go and ask.

Codex Alimentarius maintains international standards relevant to milk and milk products, including the General Standard for the Use of Dairy Terms, and standards for specific products such as milk powders, fermented milks and milkfat products. Codex is a model, though. It is not law in any country until that country adopts it.

A full compositional standard - the exact minimum fat, protein and solids-not-fat percentages required for raw drinking milk to be sold legally, as adopted by name in a specific African country - could not be obtained for this course. This module therefore prints no such figures, and you should treat anyone who quotes you universal minimums with caution.

The standard that actually binds you is set by your own country's dairy board or bureau of standards. In Kenya that means the Kenya Dairy Board and the Kenya Bureau of Standards; every country has an equivalent dairy authority and standards body. They set and enforce the sale-grade standard you must meet, and it may or may not follow Codex exactly.

So the practical instruction that ends this module is a set of questions to go and ask, and to write the answers down with a date, because standards get revised:

1. What compositional standard must my milk meet to be sold legally, and who sets it?
2. Which tests does my buyer run, and what result causes a rejection or a price cut?
3. Does my buyer screen for antibiotic residues, and with what?
4. Is there any herd testing programme for brucellosis or bovine TB that I can join or am required to join?

A farmer who knows the answers to those four questions is in a completely different position from one who simply delivers milk and hopes. And no book, this one included, can answer them for you, because the answers are set locally and they change.