Exhaust Smoke Colours: What Each One Is Burning

An exhaust pipe is the only diagnostic instrument on a used car that a seller cannot reset, wipe or reprint. Whatever is being burned in the cylinders leaves through it, and it leaves in a colour. That makes the tailpipe the cheapest test available to a buyer standing in a stranger’s driveway with no tools, no lift and no service history worth the paper it is on — provided the engine is cold when you get there, which is the whole argument of this page and the reason the last third of it is about arranging a viewing rather than about engines.

The colour is the diagnosis. White smoke, blue smoke and an engine that burns oil are not three separate questions with three separate answers; they are one question — which fluid is getting into the combustion chamber — asked from three directions. Coolant burns white. Oil burns blue. Excess fuel burns black. Each of those answers sits at a completely different price, and the gap between the cheapest and the dearest is the difference between a service item and a decision about whether the car is worth keeping.

The figure above sets each colour against the fluid behind it and against what it costs to guess wrong, because guessing wrong in either direction is expensive. Read a plume of cold-morning condensation as a head gasket and you walk away from a sound car, or accept a discount you did not need to negotiate. Read genuine coolant smoke as condensation and you buy an engine that is already leaking into its own cylinders. The rest of this page is about telling those apart with your eyes and your nose, in the order you will actually encounter them.

The white vapour that is not a fault

Start here, because this is the common case and dismissing it properly saves everyone a great deal of worry.

An exhaust system spends most of its life cold. The gas arriving in it is hot and wet, and when that gas meets cold steel the water in it condenses on the pipe walls exactly the way breath fogs a window. Some of it runs out of the tailpipe as liquid — the drip that puts a wet patch on the ground under a parked car — and the rest leaves as visible vapour. Once the system is hot enough that condensation can no longer form on the inside of the pipes, the vapour stops. On a mild day that takes a couple of minutes. On a freezing one it takes considerably longer, and on a car making only short journeys the exhaust may never get properly hot at all, which is why the same car looks fine in July and alarming in January.

Four things separate condensation from anything worth worrying about, and none of them requires a tool:

The corollary is worth stating plainly for anybody who came here in a mild panic: a car that steams on a cold start, clears within a few minutes, has a stable coolant level and does not smell sweet is behaving normally. Weather explains it. So does a car that has been doing nothing but school runs.

Thick white smoke that stays: coolant in the cylinder

Now the other white. This is coolant reaching a combustion chamber and being burned along with the fuel, and it is the answer nobody wants.

Approached from the tailpipe rather than from the engine bay, it looks and behaves differently from condensation in ways that are obvious side by side and harder alone. The plume is dense and opaque rather than translucent. It has volume, and it keeps that volume for several car lengths. It is still there when the engine has been running for a quarter of an hour and everything under the bonnet is fully hot. And it carries the smell — glycol coolant is sweet, and once the association is made you cannot un-make it.

There is a second signature that is genuinely useful at a viewing and that almost nobody looks for. Leave the engine idling for several minutes, then ask for a hard blip of the throttle. Coolant seeping into a cylinder at idle accumulates there faster than it is burned, and the sudden increase in airflow clears the lot in one dense white cloud that is out of all proportion to anything the car did before it. A car that idles almost cleanly and then produces a startling puff when revved has told you something a steady observation would have missed.

What is actually broken, and why the pipe cannot tell you

Coolant reaches a combustion chamber through a failure of the seal between the water jacket and the cylinder. That is usually the head gasket, sometimes a crack in the cylinder head itself, occasionally a crack in the block, and on some engines a failure of an internal passage or a liner seal. Those four have wildly different consequences. A gasket is a serious job but a routine one, with a known part and known labour. A cracked head means machining or a replacement head. A crack in the block is generally the end of that engine as an economic proposition.

From behind the car you cannot distinguish them, and it is important to be honest about that rather than pretending the smoke is more specific than it is. What the white plume establishes is that coolant is entering a cylinder. Which crack is passing it is a question for a pressure test, a combustion-gas test on the coolant, a leak-down test and, if it comes to it, a stripped engine. For a buyer this matters less than it sounds, because the decision at a viewing does not depend on which of the four it is. All four are reasons to stop.

Why the gauge may be behaving perfectly

Here is the part that catches people out, and it is the reason this page approaches the subject from the exhaust rather than from the cooling system.

A breach between the water jacket and the cylinder can pass in two directions, and the two produce different symptoms. Combustion pressure pushed into the coolant is the version that makes itself felt through the cooling system: gas in the jacket, coolant displaced, temperature misbehaving. Coolant drawn or pushed the other way — into the cylinder and out through the exhaust valve — removes coolant from the system without putting anything back into it. The level falls slowly, there is no external leak anywhere to find, no bubbling to see, and the temperature gauge can read perfectly normally for a long time because there is still enough coolant to do the job.

That car looks, to a buyer doing the sensible checks, entirely healthy. The tank is a little low and the seller has an explanation. The gauge sits where it should. And the only outward evidence of what is wrong is a white plume out of the back that nobody watched, because the engine was running when the buyer arrived. If you want the cooling-system half of this diagnosis — what the gauge does, what belongs in the expansion tank, why the cap is dangerous and what a fresh coolant flush conceals — that whole side of it is set out in our page on an engine that overheats. This page stays behind the car.

Two smaller white-smoke causes worth knowing

Not every white plume is a gasket. On a car with an automatic transmission of a certain age, a failed vacuum modulator can draw transmission fluid into the intake, and that burns as a whitish-grey smoke with a sharp, distinctive smell that is nothing like coolant. On an older vehicle, a failing brake booster diaphragm can pull brake fluid the same way, with the same result. Both are far less common than a gasket and both are cheaper. Both are also worth mentioning to whoever inspects the car, because a mechanic told only that it makes white smoke will start with the expensive hypothesis.

Blue smoke: the engine is burning its own oil

Blue-grey smoke means engine oil is getting into a combustion chamber and being burned. It looks thinner and wispier than coolant smoke, hangs in the air in a way that catches the light, and smells acrid and oily rather than sweet. It also leaves evidence: a fine dark film that settles on the rear bumper and the back of the car, and a wet, greasy deposit inside the tailpipe rather than a dry sooty one.

Colour perception is a real obstacle here, and it is worth dealing with before the diagnosis. Blue exhaust smoke is not blue the way a blue car is blue. It reads as a grey haze with a blue cast, it is nearly invisible against a bright sky or a dark hedge, and it shows best against a pale wall in indirect light. Position yourself so the plume passes in front of something light-coloured before you decide there is nothing there. More than one buyer has declared a car clean because they were looking at it silhouetted against the sun.

When the blue appears is most of the diagnosis

Oil can reach a cylinder from above, from below or from the turbocharger, and each route leaks under different conditions. That is what makes timing so informative: the same quantity of smoke at three different moments in the drive cycle means three different repairs at three very different prices.

The turbocharged case deserves its own note because it is the one most often misread. A turbo is oil-fed and spins in plain bearings, with seals at each end holding oil away from the compressor and the turbine. When the turbine-end seal starts to pass, oil goes straight into the exhaust and burns, and it does it most obviously on the overrun. When the compressor-end seal goes, oil is delivered into the intake tract instead, coats the intercooler and is then burned in the cylinders — which produces the same blue smoke by a completely different path, along with an oily residue in the intake pipework that is visible if anybody takes a hose off. On any turbocharged used car, blue on the overrun should be treated as a turbo question until proven otherwise, because turbos are not cheap and a failing one can take the engine with it if the compressor wheel lets go.

One more source that is not the engine

A blocked or failed positive crankcase ventilation system produces oil smoke without anything inside the engine being worn out. The system exists to draw combustion gases that have blown past the rings out of the crankcase and feed them back into the intake to be burned. When the valve or the hose blocks, pressure builds in the case, and that pressure pushes oil past seals that were perfectly capable of holding it — sometimes out through the dipstick tube or the filler cap, sometimes into the intake in quantity. It is one of the cheapest parts in the engine bay and it can imitate a worn engine convincingly. It belongs at the top of the list to eliminate, not the bottom.

Black smoke: too much fuel, or not enough air

Black smoke is unburnt fuel — soot, essentially — and it means the mixture reaching the cylinders is too rich. Either more fuel is going in than should be, or less air is arriving than the engine expects. It smells strongly of raw fuel, it comes with poor economy that the owner will usually have noticed, and on a petrol car it is very often accompanied by a warning lamp, because the engine management system can see the mixture going wrong through its own oxygen sensors and will log it.

That last point makes black smoke the one colour on this page with a paper trail. A rich-running condition sets fuel trim codes and frequently misfire codes alongside them, and those live in the module’s memory whether or not a lamp is currently lit. The page on what a check engine light is telling you explains why a stored code identifies a circuit and not a culprit — a distinction with teeth here, because a code pointing at an oxygen sensor is thrown by an engine running rich at least as readily as by a faulty sensor.

The diesel case carries a specific buying warning that the petrol case does not. Removing or defeating a diesel particulate filter, or remapping an engine to over-fuel it, is tampering with federally required emissions equipment. It is unlawful, it will fail an emissions inspection in any jurisdiction that runs one, and putting the vehicle back to standard afterwards is expensive because the parts that were removed have to be bought again. None of that is visible from the driver’s seat. Some of it is visible from underneath, and all of it is a question worth asking outright before money changes hands. If the answer is evasive on a diesel pickup that makes a lot of smoke and a lot of noise, take the evasion as the answer.

Oil consumption with nothing visible at all

A car can use oil steadily and show you nothing. This is common, it confuses people, and it does not mean the dipstick is lying.

There are two reasons. The first is that a working catalytic converter deals with a good deal of what passes through it, so oil entering the exhaust at a modest rate can be substantially consumed before it reaches the open air. The second is simply that there is a threshold below which burning oil does not make a visible plume in daylight, and an engine can sit under that threshold for years while still needing topping up between services. This page will not tell you how much oil is too much in quarts or miles, because that figure varies by engine to a degree that makes any single number actively misleading, and because the manufacturer’s own published allowance for some engines is higher than most owners would guess.

What you can do is establish the shape of it. Where oil goes when it leaves the sump is a short list: past the rings, past the valve stem seals, through the turbo, through the crankcase ventilation system into the intake, or out of the engine entirely as a leak. The first four are consumption. The last one is not, and telling them apart is the next section, because they are found in different places and they cost different amounts.

The buyer’s version of this check is documentary rather than mechanical. Ask the seller how often they top it up between oil changes, and listen to how quickly the answer comes. Somebody who genuinely tops up a car knows the answer instantly, because it is a habit with a rhythm — a bottle in the boot, a check every few weeks. Somebody who says the car has never used a drop, on a high-mileage engine, is either fortunate, not looking, or telling you what they think you want to hear. Then read that answer against the receipts. The page on oil change intervals covers how to read the cadence of a service folder rather than the intervals printed in it, which is the skill that makes the difference here: an engine that has been fed on time for its whole life and one that has been ignored do not wear at the same rate, and the folder is the only record of which one you are looking at.

Burning it or leaking it

Two cars both need a top-up every month. One is burning oil in the cylinders. The other is dropping it on the ground. The bills are different, the evidence is in different places, and the confusion between them is the single most common misreading of this whole subject.

The reason people conflate them is that a leak can produce smoke too. Oil escaping onto a hot exhaust manifold burns off in a haze that rises through the engine bay, curls out around the edges of the bonnet, and is often at its most obvious after you stop — because with no airflow the heat soaks and the residue smokes. It smells sharp and burnt. Crucially, it does not come out of the tailpipe. If the smoke is around the bonnet and the smell is strongest at the front of the car, you are looking at a leak, not at an engine burning oil internally.

Press the distinction at the viewing, because the two answers are priced nothing like each other. Sealing surfaces high on the engine are generally reachable, so the labour is short and the invoice tracks the part. Seals buried at the back are the reverse: the component may cost very little while getting a spanner near it means taking the drivetrain apart. Consumption sits somewhere else again, because nothing is escaping and the work, when it becomes necessary, is inside the cylinders. A seller describing the car as having “a slight weep” may mean any of the three, and the phrase distinguishes none of them.

Two ordinary, unsinister things wipe that evidence out. A steam-cleaned engine bay takes the film, the dirt pattern and the trail with it in a single afternoon, and most people who do that before a sale simply want the car to look cared for. An undertray refitted after a service hides the whole underside as a side effect of being an undertray. Neither proves anything either way. What both do is remove the evidence you would otherwise have had for nothing, which then has to be bought back by putting the car on a ramp — one of several reasons a proper pre-purchase inspection earns its fee on any engine with a vague oil story.

Engines that use oil because of how they were built

Not every oil-consuming engine is worn out, and treating consumption as automatic evidence of neglect will lead you to reject sound cars.

Design choices drive it. Wider piston ring clearances buy power and durability under sustained high load at the cost of a little more oil past the rings. Some high-performance engines are built that way deliberately. Direct injection shifts things as well: with no fuel spray crossing the back of the intake valve to keep it clean, whatever oil mist the crankcase breather feeds into the tract settles and hardens on the valve and port instead. Turbocharging adds a component that is oil-fed by design and that ages. And on any given engine family there may be a particular ring pack, oil-control ring or seal specification that turned out to be marginal in service and was revised later in production.

Where that leaves a buyer is here. “They all use a bit of oil” is sometimes a true statement about a specific engine and sometimes a sentence that is doing a great deal of work to cover a tired one. You cannot tell which from the sentence. What you can do is find out whether the engine in front of you has a documented consumption characteristic, whether the manufacturer published a measurement procedure for it, and whether this particular car was ever put through that procedure by a dealer. That last one is the useful question, because a consumption test performed under warranty leaves a record, and a record is the difference between a known quantity and a hopeful assumption. If the seller’s account of the car’s past is thin, you can run a VIN history check on it before you decide how much of that account to believe.

Reading the oil and the coolant against each other

When coolant and oil mix, each one carries evidence of the other, and checking both takes about a minute with the engine cold. The checks are worth doing. They are also weaker than their reputation, and the weakness is the part most people are never told.

The oil side

The famous check is the pale, thick emulsion sometimes found under the oil filler cap — mayonnaise, in the usual shorthand. It forms when water and oil are churned together, and coolant entering the sump will certainly produce it. So will something wholly benign: moisture condensing inside an engine that is switched off before it has run long enough or hot enough to boil that moisture away again, which is the ordinary state of a car making short trips in cold weather. The filler cap is the coldest, most exposed metal in the vicinity, so it is precisely where harmless condensate collects first.

What tells the two apart is spread and volume. Deposit confined to the underside of the cap, with clean oil on the dipstick, is the short-journey pattern. Emulsion on the dipstick itself, or a milky sheen through the oil, or a sump level that has risen rather than fallen, is contamination. A rising oil level on a car that has not been topped up is one of the more alarming findings available from a dipstick, because something is being added to the sump and the candidates are coolant and fuel.

The coolant side

Look into the expansion tank in good light and look at the surface rather than the level. An oily film sitting on top, a rainbow sheen, a brown or tan scum ring at the fluid line, or coolant that has gone opaque and sludgy are all signs that oil has crossed into the water jacket. Feel the underside of the cap and the inside of the neck too; deposits collect there.

One useful distinction that gets missed: oil can reach the coolant without any gasket being involved. Many engines run an oil cooler with coolant on one side and oil on the other, and a failed seal or a corroded core in that cooler puts the two fluids together with the cylinder head entirely innocent. Some automatic transmissions are cooled the same way through the radiator, so transmission fluid can appear in the coolant by the same route. Both are considerably cheaper repairs than a head gasket, and both produce a contaminated coolant that would otherwise condemn the engine. It is a distinction worth raising with whoever diagnoses the car rather than accepting the first conclusion.

The complaint file, and why smoke never appears in it

The constraint has to come before the figures on this subject, because it is a tight one. Reports to the federal safety regulator are sorted by the system involved, never by what the driver saw, and there is no field anywhere in the archive for the colour of an exhaust plume. So nothing below is a count of cars that smoked.

Two of the fifteen groups in that pull concern the engine. The one filed simply as engine carries 200 readings, its median sitting at 58,000 miles; the separate engine and engine cooling group carries 97 of them, median 98,000, against 60,000 across the file as a whole. Those are filing categories, not diagnoses. A report about blue smoke, a report about a timing chain and a report about stalling all land in the same bucket, and nothing separates them afterwards. So this page cannot tell you how many owners reported an engine that smoked, at what mileage, or in what colour, and any figure claiming otherwise is not coming from here.

What the archive is genuinely good for on this subject is a point about timing that has nothing to do with categories. Every row in it is the end of a sequence, not the start of one. Something went wrong, the owner lived with it, it got worse, and eventually it crossed the threshold at which a person sits down and files a report with a government agency. Whatever the mileage attached to that report, the first observable evidence existed earlier — often much earlier, and often at the tailpipe. Oil past a set of valve stem seals is visible on a cold morning for years before it becomes a complaint. Coolant entering a cylinder announces itself in the exhaust before it announces itself anywhere on the dashboard.

That gap is the whole argument for the section that follows. The mileages in the file are where problems get escalated. The exhaust is where they get noticed, and the price you pay for a car is fixed at the noticing end, not the escalating one. It is also worth saying that both engine medians fall inside the range at which perfectly ordinary used cars are advertised, so no odometer reading you are likely to encounter puts a car outside the window in which this matters.

Engines in the recall record, and what is deliberately absent from it

Recalls are a second archive and a different order of evidence. One is testimony from an owner about a car they drive; the other is a manufacturer accepting, in writing and at its own expense, that a population of vehicles left the factory with a defect. Keep them apart.

Engine is a small family in that archive and an oddly shaped one. It holds five campaigns raised by three separate manufacturers, and it is one of the few families in the archive with no parts breakdown beneath it — every one of the five was filed against the engine as a whole, with no component named inside it. The other families in that position are all single-campaign entries, which makes engine much the largest of them. That is less a gap in the data than a description of what an engine recall tends to be: a fault in something structural or something that turns, rather than in a part with a tidy name. The neighbouring family, engine and engine cooling, is larger at eight campaigns from four manufacturers, and its named parts run to the cooling circuit, a pump and an exhaust gas recirculation valve.

The absence is what matters most for this page. There is no recall for burning oil, and there will not be one, because a campaign exists to address an unreasonable risk to safety — fire, a stall in traffic, a sudden loss of motive power — and an engine that quietly consumes oil is a durability and warranty question rather than a safety one. Manufacturers deal with consumption through service bulletins, revised parts, extended coverage and goodwill, and none of that appears in the public recall lookup or answers to a VIN the way a campaign does. So a clean recall check on a car that smokes tells you exactly nothing about the smoke. Run it anyway, because an open campaign is free work no matter how many times the car has changed hands, but do not let a clean result stand in for an inspection.

At a viewing: watch the tailpipe yourself

This is the section the rest of the page exists to support. Everything above describes evidence that appears for a few seconds, once a day, at an event you have to be present for. Being present for it is a logistical problem more than a mechanical one, and it is solved before you arrive.

The general case for insisting on a cold start — what the cranking speed, the time to fire and the settling idle each reveal — is made in full in our guide to a car that refuses to start, and is not repeated here. What follows is narrower and concerns one thing only: getting yourself behind the car, in the right position, at the moment the engine first fires, and knowing what you are looking at when it does.

Arrange it, then arrive before you are expected

Put the request in the message that fixes the viewing, in plain terms and with the reason attached, because a request with a reason is much harder to forget than a bare instruction: please leave the car standing and do not start it or move it before I get there, as I want to watch it started from cold. Most sellers simply agree. Some will have moved it anyway, which is why the second half of this matters.

Arrive a few minutes early rather than exactly on time, and do your first check before you knock on the door. Everyone knows to feel the bonnet, including sellers. Fewer people think about the exhaust, and the exhaust is the better test for the specific question this page asks. Hold the back of your hand near the tailpipe — near, not against it, because a hot exhaust burns immediately and silently. Look at the ground under the pipe as well. A car that has been run in the last hour on a cool morning will very often have left a damp patch there from the condensate that dripped out of it during warm-up. A dry pipe over a dry patch on a damp morning is consistent with a car that genuinely has not been started. A wet patch under a car described as untouched is a piece of information.

Where to stand, and who turns the key

You get one cold start. You cannot watch the dashboard and the tailpipe at the same time, and you have to choose.

Choose the tailpipe. The warning-lamp sweep is repeatable — you can switch the ignition on and off as many times as you like later in the visit and see exactly the same sequence. The cold-start plume is not repeatable. It happens once, it is over in well under a minute, and after that the car will not produce it again until tomorrow morning. So on the first start, hand the key back and ask the seller to start it while you stand at the back. Take the key yourself for the second start, later, when it costs you nothing.

Stand slightly to one side of the pipe rather than directly behind it, a few feet back, and position yourself so the plume will pass in front of something pale — a light wall, a garage door, a white van, the sky if it is overcast rather than bright. Against a dark background you will miss a blue haze entirely. If the car has two tailpipes, watch both; a single failing cylinder bank can smoke through one and not the other, which is itself diagnostic.

The first thirty seconds

Watch in three phases, because the shape of the plume over time carries more than any single moment of it.

The blip, and the overrun

Two deliberate tests are worth asking for, and both are reasonable requests that a straightforward seller will not mind.

The first is at standstill. With the engine fully warm and having idled for several minutes, ask for a firm blip of the throttle while you watch from behind. You are looking for a cloud disproportionate to the action: dense white says coolant has been pooling in a cylinder, blue says oil has been accumulating past seals or in the turbo, black says over-fuelling. A clean engine produces almost nothing.

The second is on the road, and it is the test most buyers never manage because it requires a second pair of eyes. Smoke under load and smoke on the overrun are the two conditions that expose piston rings and turbo seals, and you cannot see either from the driver’s seat. Take somebody with you and have them follow in a second car for a few minutes of the test drive, through one hard acceleration up a gradient and one long closed-throttle descent. If that is not possible, use a straight stretch, accelerate hard, then lift off completely and watch the mirror — imperfect, but far better than nothing, and it will catch a substantial plume even if it misses a faint one.

The tailpipe itself

When the car has cooled, wipe a finger around the inside of the tailpipe as far in as you can reach. Be aware that on many cars the visible chrome or stainless trim is not the exhaust at all, merely a cosmetic sleeve over the end of it, so reach past the trim to the actual pipe.

Dry black soot that brushes off your finger is normal for a petrol engine and tells you nothing bad. A wet, greasy black that smears is oil, and it is one of the most direct pieces of evidence of consumption available without tools. A white or pale grey chalky deposit is worth noting alongside any white plume. And an interior that is spotlessly clean on a car with real mileage is itself a small finding: exhaust pipes accumulate deposits, so an unnaturally clean one has been cleaned, and it is fair to wonder why that was on the list of things to do before a viewing.

When you arrive and the engine is already warm

It happens often enough that you should decide in advance how you will handle it, because the wrong move is either to shrug and carry on or to make an accusation you cannot support.

Say it out loud, without heat, as a fact rather than a complaint: this one is already warm, and I had asked to watch it from cold — how do we fix that? Then stop talking and let the seller answer. Ordinary explanations arrive instantly and without elaboration: it lives in a garage and had to come out, it was parked at a relative’s house, they forgot. The reply that should raise your guard is the one that sets about persuading you the cold start is unimportant, since no one who has ever sold a vehicle actually believes that.

Then price the situation rather than argue about it. Best case, you return on another day and turn the key yourself — one extra journey in exchange for the single piece of evidence this whole page is built on. Failing that, leave the engine untouched for the rest of your visit, work through the documents, the bodywork, the tyres and everything under the bonnet while it sits, and fire it up last of all, after the longest cooling period you can manufacture. That is not a genuine cold start and will not show you valve seals properly, but it will still catch an engine that pools oil while standing. And failing that, you buy without the evidence and let your offer reflect the risk, which is a perfectly defensible thing to do provided it is a decision and not something you drifted into.

What does not exist is a version of this where nothing has been lost. If the cold start is gone, pay for the equivalent: a compression or leak-down test with a borescope down the bores recovers most of what the exhaust would have shown you, and neither is exotic. Do the documentary work at the same time, since a vehicle that has moved between owners in quick succession, or one whose past the seller has skated over, changes how much a single tidy start is worth — and reading what the VIN record actually says about the car takes less time than a lap of the bodywork.

One final calibration, since people agonise over this. Missing the cold start once means very little; diaries collide and cars get moved. Being unable to arrange it on a second attempt, for a vehicle the advertisement describes as needing nothing, is a different signal altogether. You do not owe anybody an explanation of what you inferred. You only have to let yourself infer it.

Common questions

Is white smoke from the exhaust always a serious problem?

No, and most of the time it is not a problem at all. Thin white vapour that appears on a cold start, thins as it leaves the pipe, disperses within a few feet of the car and stops once the exhaust is hot is condensation, and it is more pronounced in cold, damp weather and on cars that only make short journeys. The version that matters is thicker, holds its shape as it drifts away, carries a faintly sweet smell and is still there after the engine has been running for a quarter of an hour. That one is coolant reaching a cylinder, and it means a head gasket, a cracked head or a cracked block.

What does blue smoke from the exhaust mean?

Engine oil is getting into the combustion chamber and being burned. When it appears tells you which route the oil is taking. A puff on the first firing from cold that clears within seconds and does not return that day is the signature of worn valve stem seals, because oil drains down the stems overnight and burns off in one go. A haze that builds under sustained acceleration or on a long climb points at piston rings and bore wear, because load is what forces oil past a ring pack that no longer seals. A cloud when you get back on the throttle after lifting off points at valve stem seals on a naturally aspirated engine and at turbocharger seals on a turbo car.

My car uses oil but there is no visible smoke. What is going on?

That combination is common and does not mean the dipstick is wrong. A working catalytic converter deals with a good deal of what passes through it, and there is a rate of consumption below which nothing visible appears in daylight. The more important question is whether the oil is being burned or lost. A leak leaves outside evidence: an oily film on the engine, dirt sticking to it in downward streaks, drips on the ground, a smell of hot oil after a run, and smoke around the bonnet rather than from the tailpipe. Consumption leaves nothing outside and a greasy black deposit inside the tailpipe. A car can be doing both.

Is it normal for an engine to use some oil between changes?

For some engines, yes, and treating any consumption as evidence of a worn-out engine will make you reject sound cars. Wider piston ring clearances, direct injection, turbocharging and specific ring or seal specifications all move the figure, and some manufacturers publish an allowance considerably higher than owners expect. What that does not do is make the phrase “they all use a bit” true of the particular car in front of you. Find out whether the engine has a documented consumption characteristic, whether the manufacturer published a measurement procedure, and whether this car was ever put through it by a dealer — because that leaves a record, and a record is the difference between a known quantity and a hopeful assumption.

What causes black smoke from the exhaust?

Unburnt fuel, which means the mixture is too rich: either too much fuel is being delivered or not enough air is arriving. On a petrol engine the usual suspects are a leaking or stuck injector, a coolant temperature sensor telling the management system the engine is still cold, a mass airflow sensor under-reading, or a genuinely restricted intake. It nearly always sets fuel trim or misfire codes that remain in the module’s memory whether or not a lamp is lit. On a diesel the list is broader and includes injectors, an EGR valve stuck open, a turbo that is not making boost, and a particulate filter that is not doing its job — or that has been removed, which is unlawful tampering and expensive to reverse.

Does a clean oil filler cap mean the head gasket is fine?

Not on a car whose oil was changed recently. Emulsion under the cap builds up over time, so a fresh fill has had no opportunity to build up anything, and a spotless cap on a car serviced last week is no information rather than good news. The same logic applies to the coolant: newly flushed coolant is bright and clear by definition and cannot show you an oil film that had been sitting in it for a year. Date the fluids before you read them, look at the dipstick and the oil itself rather than only the cap, and remember that the exhaust is the one check that no fluid change erases.

Should I buy a used car that smokes on start-up?

It depends entirely on the colour and on what happens next, which is why seeing it yourself matters so much. A brief blue-grey puff on the first firing that clears and does not return is a valve stem seal signature; on its own that is not grounds for walking away, but it is grounds for pricing the car as one that wants work on the top end. Thick white that persists once the engine is fully warm, with a sweetish smell, is coolant entering a cylinder and is a reason to stop. If the engine was already running when you arrived, you have not seen any of this — so either arrange a second viewing and start it yourself, or pay for a compression or leak-down test and a borescope before any money moves.