'It's usually pretty dark, smelly, uncomfortable and physically demanding. The days are long and you come home filthy and stinking of burnt plastic. There's nothing glamorous about it. But it is fascinating.'

Niamh Nic Daéid, fire scene investigator

Sunday, 2 September 1666. A family servant coughs himself awake in Pudding Lane,London. Realising there's a fire in the shop below, he pounds on the bedroomdoor of his master, baker Thomas Farriner. The whole household crawl alongrooftops to safety, except for the maidservant Rose, who, paralysed with fear,perishes in the blaze.

Soon flames begin licking the walls of neighbouring houses and the Lord Mayor,Sir Thomas Bloodworth, is called upon to authorise the firemen to pullbuildings down to stop the fire spreading. Bloodworth is angry at having hissleep disturbed and ignores the firemen's urgent demands for drastic action.'Pish!' he says. 'A woman could piss it out.' And leaves the scene.

In the middle of the morning, diarist Samuel Pepys experiences 'the wind mightyhigh and driving [the fire] into the city, and everything, after so long adrought, proving combustible, even the very stones of churches'. By theafternoon London is in the grips of a firestorm, roaring through 'warehouses ofoyle and wines and Brandy', wooden houses, thatched roofs, pitch, fabrics,fats, coal, gun powder – all the flammable material of life in the seventeenthcentury. The immense heat of the blaze makes escaping gases rapidly expand andrise, sucking in fresh air at gale force speeds, feeding the inferno with yetmore oxygen. The Great Fire has created its own weather system.

When the fire abates four days later, it has destroyed most of the medievalCity of London, including more than 13,000 houses, 87 churches and St Paul'sCathedral. Roughly 70,000 of the city's population of 80,000 are suddenlyhomeless.

The ashes were still warm when the conspiracy theories started. Most Londonerscould not bring themselves to believe the fire was an accident. There were toomany coincidences for that; it started among tightly packed wooden buildings,while everyone slept, on the one day of the week when the streets were empty ofhelping hands, when a gale was blowing, and the Thames lay at low tide.

Rumours of foul play were rife. A surgeon, Thomas Middleton, had stood at thetop of a church steeple and watched as fires seemed to break out in severaldistinct and distant places at once. 'These and such like observations begat inme a persuasion that the fire was maintained by design,' he wrote.

Foreigners in particular were suspected, with one Frenchman beaten nearly todeath in Moorfields for carrying 'balls of fire' in a box; they turned out tobe tennis balls. Poems and songs expressed the bewilderment at the fire'sorigin and cause:

origin and cause:ed out to be tennis balls.

Whether from Hell, France, Rome, or Amsterdam.

Anon., 'A Poem on the Burning of London' (1667)

The desire to know the truth started at the very top. Charles II had lost moreproperty in the fire than anyone else. The king empowered parliament to set upa committee of inquiry into the fire. Scores of eyewitnesses came forward.Several said they had seen people throwing fireballs, or confessed to throwingthem themselves. One Edward Taylor said that on Saturday night he went with hisDutch uncle to Pudding Lane, found the window of Thomas Farriner's bakery open,and threw in 'two fireballs made of gunpowder and brimstone'. But as EdwardTaylor was only ten years old his account was dismissed. Robert Hubert, thesimpleminded son of a French watchmaker, confessed to having started the fire.No one really believed him but because he insisted the jury found him guilty,and he went to the gallows at Tyburn.

One member of the parliamentary committee, Sir Thomas Osborne, wrote that 'allthe allegations are very frivolous, and people are generally satisfied that thefire was accidental'. In the end, the committee decided that the dreadfulconflagration was caused by 'the hand of God, a great Wind, and a very dryseason'.

It's not surprising that the committee arrived at so unsatisfactory aconclusion. For investigators to evaluate complex fire scenes, they need tounderstand how fire works. Back in the seventeenth century, the scientificknowledge was woefully insufficient. It wasn't until 1861, when Michael Faradayput his lectures on fire into a book, that such understanding became readilyavailable to a wide audience. The Chemical History of a Candle was thepublished version of six lectures he delivered for a young audience, but it isstill regarded as a key text on the subject. Faraday used the candle as asymbol to illuminate the general nature of combustion. In one key lecture hesnuffed a candle out by putting a jar over it. 'Air is absolutely necessary forcombustion,' he explained. 'And, what is more, I must have you understand thatfresh air is necessary.' By 'fresh air' he meant 'oxygen'.

Faraday was an early expert witness, taking with him his findings from thelaboratory – sometimes quite literally. In 1819 the owners of a sugarhousedestroyed by fire in Whitechapel, London, sued their insurance company who hadrefused to pay out £15,000 compensation. The case turned on whether or not a newlydeveloped process involving heated whale oil – which the owners had startedusing at the factory without the insurers' knowledge – had made the fire moreor less likely. Before testifying, Faraday performed experiments on whale oil,heating it to 200°C to demonstrate that 'all the vapours of the oil, exceptwater, are more inflammable than the oil itself'. In court a member of the jurydid not believe him, so Faraday set fire to some of the oil's distilled vapours(naptha) which he had brought with him in a vial, 'a most offensive smell beingat the same time perceived throughout the court'.

Michael Faraday, whose 1861 book The Chemical History of a Candle paved the wayfor modern fire scene investigators

Faraday's most important forensic investigation was into an explosion atHaswell Colliery, County Durham, which killed ninety-five men and boys in 1844.The blast occurred at a time of industrial unrest in the Durham coalfield. Thelawyer acting for the grieving relatives petitioned the Prime Minister, RobertPeel, to send government representatives to the inquest. Faraday was amongthose sent.

The team spent a day visiting the mine, investigating in particular its airflows. At one point, Faraday realised he was sitting on a keg of gunpowder neara naked candle flame. He leapt to his feet and 'expostulated with them fortheir carelessness'. The jury reached a verdict of accidental death, with whichFaraday agreed. But the team submitted a report on their return to Londonnoting that coal dust had played a major part in the explosion, andrecommending that the ventilation be improved. The mine owners objected,because of the costs of improvement. The risk was ignored for sixty years,until a similar explosion in 1913 led to the death of 440 miners at SenghenyddColliery in Wales – the worst mining disaster in UK history.

In the twentieth century, the Fire Service and the scientific communitydeveloped fire scene investigation in tandem, encouraged by governments whowanted to know how many fires there were, their origins and their causes. Inthe 1960s and 1970s investigations became more rigorously scientific: protocolswere adopted; new instruments enabled complex chemical mixtures such as petrolto be identified at fire scenes; and experts in the field began to emerge.Partly as a result of this increased understanding, it is now rare for a fireor an explosion – which is essentially an expedited fire – to cause suchhorrendous loss of life in peacetime. But when they do, they leave an indelibleimpression on those who investigate them.

Among those who became the new experts in fire investigation were an Irishhusband and wife. Their daughter, forensic chemist Niamh Nic Daéid of theUniversity of Dundee, has continued their legacy, searching for the truth amidscenes of terrible destruction. Niamh explains. 'I have a legacy in forensicsciences, if you like, because my parents were both independent fireinvestigators, and indeed my mother still does fire scene investigation, so Igrew up with it. Myself and my brother used to make our pocket money bysticking Mum and Dad's fire pictures into reports – for five pence a picture.As you can imagine, the conversation around the dinner table was always aboutfires.'

Whether fire obliterates someone's property or their dearest relative, theinvestigator works at the point of collision between nature's most violentforce and the human world it wrecks. I was forcibly reminded of that when Iasked Niamh about fires that had particularly affected her. The first words outof her mouth were: 'The Stardust disco fire.'

I was asleep in my bed in Derbyshire in the early hours of Valentine's Day,1981. I was a young journalist, based in the northern newsroom of a nationalSunday paper. I'd never covered a major disaster but I knew that was about tochange when a ringing phone woke me in the small hours of the morning. Thefamiliar gruff voice of my news editor said, 'There's been a major fatal firein a Dublin disco. It's looking like dozens dead. You're on the seven o'clockplane.'

By the time I got to Manchester Airport, the radio had confirmed what I'dalready been told. A massive fire. A horrifying death toll of young peoplewho'd set off for a fun night out and who wouldn't be going home. Inside theairport, journalists and photographers milled around, looking for colleagues sothey could hunker down in their own little scrums and divide up the tasks onthe ground at the other end.

My own team – three other reporters and two pic men – made our way to a cornerof the bar. A double whisky was set in front of me. Even in those hard-drinkingnewspaper days, I wasn't accustomed to starting my day that way. 'Drink it,'one of my colleagues insisted. 'Trust me, you're going to need it beforetoday's over.'

He was right. When we landed at Dublin, our Irish staff reporter gave us thegrim news. More than forty dead. Because I'm a woman, because I was consideredto be good with the grieving but also hard-headed enough to get what I'd comefor, I was assigned to the death knocks – visiting the bereft families so wecould flesh out our story with poignant quotes and photographs of the dead.

I spent the rest of the day on the Coolock council estate, where many of theteenagers who died in the Stardust had come from. The families were in shock,but oddly grateful that someone wanted to mark the passing of their children.I'd never spent a more harrowing day at work. And I was just a spectator. Myheart felt hollow when I imagined what the bereaved were going through.

After the first edition deadline had passed, I met up with one of my team atthe site of the fire. From the front of the building there wasn't much to seeexcept broken windows and smoke staining the upper part of the facade. Apartfrom the throat-catching stink of smoke and char, it was hard to believeforty-eight people had died and more than 240 had been injured there. It wasthe interior of the building that had been devastated by the fire; from theoutside, the only giveaway was the number of fire engines and police vehiclescrowding the roadway outside.

Niamh Nic Daéid's mother was one of those charged with finding out whathappened inside the Stardust disco that night.

The Valentine's Dance at the Stardust was supposed to be a night to rememberfor very different reasons. Eight hundred and forty-one people, mainly in theirlate teens, had handed over the £3 entrance fee that entitled them to sausageand chips and the right to dance till two in the morning, thanks to a speciallate licence.

Twenty minutes before closing, the DJ announced the winners of the best dancingprizes. A minute later, some revellers spotted smoke coming from behind aroller blind to the left of the dance floor. Most of them put it down to aspecial disco effect and kept dancing.

Behind the blind were five rows of tiered cinema seats. When some of thedancers looked behind the blind they saw some of the seats on the back rowablaze. Their polyurethane stuffing was already emitting black clouds ofextremely poisonous hydrogen cyanide. At first the fire was small andcontrollable, but it quickly grew in intensity. Employees emptied waterextinguishers into the flames – to no avail. Within five minutes, moltenplastic was dripping on to the patrons on the dance floor; part of the ceilingcollapsed on to them; and thick, toxic smoke filled the entire ballroom.

Survivors spoke of their shock at the swiftness of it all.

Fire scene investigators at the scene of the Stardust Disco Fire, in whichforty-eight people died and more than 240 were injured

When people panic they instinctively try to leave a building the same way thatthey came in, so the narrow foyer leading to the Stardust's main entrancequickly became a bottleneck. Those sprinting to the main doors found themlocked shut and it took a bouncer crucial minutes to squeeze through thedesperate crowd with the key.

But still disaster should have been averted. There were six fire exits in theStardust. But the owner, Eamon Butterly, had been worried about people openingthe doors from the outside and slipping into the venue without paying, so oneof the fire exits was locked and others had chains wrapped around them so theyappeared locked. Panic-stricken patrons tried, and eventually managed, to kickthese doors open. Another fire exit had tables and seats stacked on either sideof it; yet another had plastic skips blocking it.

At 1.45 a.m., when the ceiling in the ballroom collapsed and the electricitycut out, around 500 people were still inside. The blistering flames were theironly source of light. The Adam and the Ants record that had been playing wasreplaced by terrified screams. Within nine minutes of the fire being spotted,everything in the Stardust was ablaze – seats, walls, ceiling, floor, tables,even metal ashtrays.

In the mayhem, some people fled into the toilets. Six weeks before the disco,Butterly had heard customers were trying to smuggle alcohol in through thetoilet windows, so he'd had steel plates welded to their inside, to complementthe metal bars that were already in place outside. When fireman arrived at thescene, eleven minutes after the fire started, they attached cables to the barsand drove away at speed, but only managed to bend them. The people in thetoilets were trapped in an inferno of flame and smoke.

Everyone in the surrounding area, in the working-class communities of Artane,Kilmore and Coolock, knew someone affected by the tragedy. The whole of Irelandmourned the forty-eight killed. Five of the dead had been so badly burnt thatthey could not be identified. (In 2007 their bodies would be exhumed from acommunal grave, so DNA analysis could separate and identify them.)

At 8.35 on the morning of Valentine's Day, Detective Garda Seamus Quinninspected the gutted Stardust. He spent five hours examining the site, findingno trace of accelerants or electrical problems in the area where the fire hadfirst been spotted. He also discovered, by throwing a lit cigarette on to asimilar seat, that its non-flammable PVC covering did not catch fire. Hadsomeone slashed a seat and deliberately lit its polyurethane filling?

The British Fire Research Station carried out a full-scale reconstruction ofthe area where the fire had first been spotted at their hangar at Cardington,Bedfordshire. Investigator Bill Malhotra managed to get seats to catch lightboth by slashing them open, and by placing several sheets of newspaperunderneath them. The flames reached the very low ceiling and started meltingthe carpet tiles, causing molten drops to fall on to other seats. Within the tightspace all the seats heated up, and the boiling drops were enough to defeattheir PVC coverings. Once five seats in the back row were in flames, the row ofseats in front caught fire, too. Quinn's and Malhotra's experiments bothsuggested arson.

In June 1982, eighteen months after the blaze, the Irish government publishedthe results of a public inquiry into its origin and cause. On the question ofwhy, the report was ambiguous. 'The fire was probably caused deliberately,' itsaid at one point. At another, 'The cause of the fire is not known and maynever be known. There is no evidence of an accidental origin and equally noevidence that the fire was started deliberately.' The forensic experts who hadgiven evidence were divided. While Quinn, Malhotra and one other thought thefire was most probably caused by arson, two others wouldn't rule out anelectrical fault.

The report damned Eamon Butterly for many things, including not complying withelectrical safety standards and using locks rather than doormen to guard doors.The cost of employing extra doormen would have been £50 – just over £1 forevery life lost. On the question of the armoured toilet windows the reportsaid, 'While their primary purpose was for ventilation, it might have beenpossible for a person to get through them in an emergency.' Despite all ofthese points, the report legally exonerated Butterly from responsibility forthe fire because it had been 'probably caused by arson'. So in 1983 the statepaid Butterly compensation for malicious damages in the region of £500,000. In1985 the victims' families received an average of only £12,000 each.

The families were far less interested in money than in why their relatives hadperished. So much potential evidence had been obliterated that it seemedunlikely they would ever get an answer. But that didn't stop them trying. In2006 the Stardust Victims' Committee enlisted the help of a new set of forensicexperts in a bid to make the case for a new public inquiry. Those expertspointed out that in the reconstruction in the hangar in Cardington the fire hadtaken thirteen minutes to burn through all the seats and had never breached theroof, whereas the real thing had shot from the first seat it was seen on – at1.41 a.m. – up into the night sky within five minutes. Something didn't add up.

The experts also drew attention to various witness accounts that supported thisview. Eyewitnesses standing outside the building said they had seen flamescoming from the roof several minutes before 1.41 a.m. In the weeks leading upto Valentine's Day, Stardust employees had seen a smoke-like substance and'sparks' coming from the Lamp Room above the Main Bar, which was right by therows of blazing seats. On Valentine's Day itself, Linda Bishop and her friend hadbeen sitting below a grille in the ceiling, listening to 'Born to be Alive',when they felt a great increase in temperature. Linda looked down at the newdigital watch she'd been given for Christmas. It read '1.33'. A barman who hadfought the fire on Valentine's Day said he had 'felt a monstrous heat comingfrom the ceiling. I was positive that the fire started up in the ceiling.'

The Stardust Victims' Committee experts came to the conclusion that the burningceiling had set fire to the seats, rather than the other way around. Theybelieved that an electrical fault in the Lamp Room – which was located in theroof space and contained spot-lamps and plastic seats – had ignited theceiling. Right by the Lamp Room was a storeroom, and the experts thought theoriginal inquiry had been misled about some of its contents. Eamon Butterly'ssolicitor had provided a list of the 'approximate contents' of the storeroom,including 'bleaches, brio wax, aerosols, petrolbased waxes and polishes', butdid not mention the highly flammable 'drums of cooking oil' which were alsopresent.

Professor of Fire Dynamics Michael Delichatsios reasoned that if enough heathad come from the Lamp Room, the highly flammable contents of the storeroomwould have spontaneously combusted. This would account for the extreme speed atwhich the fire spread, raining burning plastic on to the heads of people on thedance floor, and eventually bringing the whole ceiling down. In 2009 thegovernment commissioned Senior Counsel Paul Coffey to examine the StardustVictims' Committee case for a new public inquiry. He found the 'probablydeliberate' finding of the original inquiry 'so phrased as may well give themistaken impression ... that it is established by evidence that the fire wasstarted deliberately and not a mere hypothetical explanation for the probablecause of the fire'. He recommended against a new inquiry, but suggested thegovernment change the public record to make it clear that the cause of the fireis unknown. So, twenty-seven years after the most lethal fire in Ireland'shistory, the government officially clarified the cause of the fire as unclear.Because the Lamp Room had been 'totally destroyed', 800 eyewitnesses and scoresof dispassionate forensic scientists could never know if it was the true originof the blaze. The secrets of the fire were destroyed with it. And that is oftenone of the frustrations of fire investigation.

Fire scenes vary in their complexity, but even relatively simple ones challengethe investigator, who must try to reconstruct a destructive chain of events.Let's take a typical scenario. A passer-by sees a house on fire and calls thefire brigade, who put it out. A structural engineer declares the building safeto enter, and a fire scene investigator like Niamh Nic Daéid arrives todetermine the origin of the blaze, why it happened, and how it spread.

First of all – and unusually for a forensic practitioner – Niamh may sometimesinterview eyewitnesses. Where exactly did they see the fire? Was there yellowflame and white smoke, which petrol gives off, or the thick black smoke ofburning rubber? Getting the best out of eyewitnesses is a skill. Niamh is oftentalking to people on the very edge, sometimes after the centre of their worldhas burnt down. Occasionally the fire investigators may have to 'stop theinterview and let the police know that this person might have turned into asuspect'. It is a well-known axiom that industrial fires increase when businessconditions get tough, as some firms consider the advantage of a successfulinsurance claim over a loss-making factory. Arson aside, when accidents dohappen people can be cagey. When Niamh asks employees where they were smokingbefore a fire in their office started, they usually say in the allocated area.But experience has told her that 'when it rains people tend to smoke beside theback door, where the rubbish is'.

The fossilised remains of a diatom – a single-celled organism – viewed under amicroscope

The talking over, Niamh walks around the outside of the building and letsthings sink in. Are there patterns of smoke on the walls? Which windows arebroken? Anything potentially significant in the garden, like a petrol can orcigarette ends scattered about? Then she walks through the building 'hands inpockets, not picking things up', looking for anything unusual. Now she's readyto get dirty. Outside she deals with the petrol cans and cigarette ends she sawearlier, 'photographing them in situ with a scale if possible, drawing them ona plan, packaging and labelling them appropriately'. Inside she approaches 'thebusiness end' – where the fire most likely started, moving from areas of leastto most damage in a systematic way, documenting and photographing the scene asshe progresses.

As a fire spreads from its point of origin it creates more heat, which ignitesmore material in a self-sustaining chain reaction ruled by the supply of fueland oxygen. By the time it stops burning it has often brought down ceilings andwalls, which shield things as they fall. The scene is all the more resistant tointerpretation once firefighters have directed thousands of gallons of waterinto it. 'So you've got your burnt-out shell of a house with material all overit. In order to get to the bottom of where it started, you need to de-layer it,like an archaeological dig.' Like a pathologist sawing open a ribcage toperform an autopsy, Niamh has to cause more destruction to reach her answers.She works from the area of least damage inwards because 'if the big black holein the corner is where the petrol was poured, and you march over to that andwalk around in it, you'll cross-contaminate your scene'. In extreme cases, theinvestigator will use tape to construct a grid on top of the scene, numberseach square and takes everything out in buckets to sieve for any evidence thatmight have survived.

Because fires tend to rise and spread sideways, they sometimes leave thecharred outline of a 'V' pointing at their origin. Things are less clear cutwhen an arsonist has sloshed petrol throughout a house. Thin lines of severeburning on the ground, surrounded by milder burning, can indicate a petroltrail, but flames follow the petrol's path with such speed that a single pointof origin is well nigh impossible to discern. If Niamh finds several widelydistant instances of equally bad burning, this may also indicate arson; twounconnected accidental fires beginning at the same time in a house is avanishingly rare occurrence.

Once Niamh has found the most likely origin/s of the fire, she looks forpotential sources of ignition – matches, lighters, candles; and fuel – TVs,newspapers, rubbish bins. Arsonists often leave matches behind, assuming theywill burn away to nothing. But the powdered rock in a match head contains thefossilised remains of single-cell algae called 'diatoms'. A diatom's shell ismade of silica, which is abrasive enough to help you strike the match, andtough enough to endure extremely high temperatures. Each of the 8,000 knownspecies of diatom has a unique shell structure, identifiable through amicroscope. Different brands make their matches using powdered rock fromdifferent quarries. If forensic scientists can spot the diatoms, they canidentify the match brand. Then a search of a suspect's pockets or CCTV footage fromlocal shops can provide incriminating evidence.

In her mind Niamh tries to imagine how the scene was arranged when the firebroke out. Then, as far as possible, she reconstructs it for real. Fireinvestigators don't always get this right, as Niamh once experienced in thecase of a suspicious house fire that had begun on a desk. The police asked theFire investigators to put the sooty items back on the desk in their originalpositions. When Niamh was called in to review the scene, she thought it best todo her own reconstruction and compare it to theirs.

'The other investigators had reconstructed it in a way not sustained by thephysical evidence, not noticing things like a circle where a cup had protectedthe desk from smoke. They'd put the items in the wrong place and takenphotographs which told an incorrect version of the story. With the items putback in their correct position the set of circumstances that created the firecame to life.' In 2012 Niamh ran a series of workshops relating to fire scene investigationin Scotland, which concluded that, whilst many investigators are very wellequipped for the job, '97 per cent of fires in Scotland are investigated bypersonnel who have less than a week's training in fire scene investigation'.While many of these fires are relatively straightforward to investigate, thepoint relating to appropriate training still remains. Trained fireinvestigators are critical in the correct determination of the origin and causeof a fire, and this is particularly the case 'in fire fatalities where theinvestigators have a huge obligation both to the victims and to their relativesto be able to say how individuals died in that fire.'

Mishandling evidence leads to confusion and to conflicting versions of eventsbeing presented in court. It's crucial to get it right first time, not leastbecause the clues are often so fragile. Can you get fingerprints? Can you getDNA? Can you recover information from a hard drive in a melted computer? 'Theanswer to all of those is "yes", if you have the awareness not to go clumpingaround damaging material.'

Treading lightly is not easy for Niamh in her heavy-duty steel-toecapped boots,hard hat and protective overall. The scenes that she enters can contain liveelectrical hazards, jagged glass, partially collapsed walls. 'It's usuallypretty dark, smelly, uncomfortable and physically demanding. The days are longand you come home absolutely filthy and stinking of burnt plastic. There'snothing glamorous about it. But it is fascinating.'

At the suspected point of origin, Niamh collects the debris and sifts throughit by hand. 'You would be astonished at what survives. Fires are destructivethings, but they generally leave quite a lot of material behind. Things likebuttons, lighters, bottles, beer cans, anything metal, survive relatively well.Plastic materials can be melted on one side but fine on another. So you mightbe able to lift a fingerprint from the underside of a TV remote control.'

Electricity can be your friend in the fire scene, and can provide corroboratingphysical evidence relating to the cause, origin or spread of the fire. Fireinvestigators like Niamh crawl around in the muck armed with pliers, followingcables as if they were Ariadne's thread guiding her through the labyrinth.'Many scene investigators don't see the value of the electrical circuitry. It'svery laborious and time-consuming work, but enormously useful because it givessolid physical evidence, compared to burn patterns which can be interpretedmore subjectively.'

On the wall of her office Niamh has two photographs of a 12-storey building byPiccadilly Tube station in London. The top seven storeys were wrecked by fireto the tune of £12 million pounds' worth of damage. When the investigatorsfirst arrived at that scene they spoke to a cleaner who reported that she'dspotted the fire when it was still small, in the lighting system of oneparticular floor. That gave the investigators a pointer, but finding the exactorigin of such a severe fire was nevertheless still daunting. Niamh spent twodays in the building with her colleagues before they finally tracked it down toan electrical fault within a water cooler. 'It was a really interesting firebecause it involved a lot of use of the electrical system to corroborate thearea of origin. So it's dear to my heart, which is why I've got a picture of iton the wall.'

Some fires begin with electrical faults. But others have less innocent origins.Fire scene investigators will often bring in sniffer dogs, whose sense of smellis 200 times more sensitive than a human's, to find ignitable liquidaccelerants like petrol, paraffin and white spirit. There are about twentyhydrocarbon dog teams in the UK, many of whom wear little boots to protecttheir paws (and to protect the scene from contamination). 'I've seen them inaction and, boy, are they good. They just sit down and indicate when they smellsomething,' says Niamh.

Once a dog has identified the presence of a hydrocarbon, the fire investigatorstarts to bag the evidence. Because plastic bags react with the hydrocarbons insubstances like petrol, they put suspicious material into nylon bags and takesit back to the forensic science lab for analysis. If the material is somethinglike a piece of carpet, the investigator tries to take a separate, unburntpiece from the scene, for comparison. In the laboratory forensic chemists willanalyse the fire debris submitted. They use various techniques to extractpossible chemical accelerants, including 'headspace extraction'. The mostcommon way of doing this involves placing the material in a closed containerand heating it to allow vapours to rise off it. These are then collected usingan absorbent material, and extracted using a chemical solvent. From this vapourthe forensic chemist tries to identify particular compounds, usually using gaschromatography. This is a fairly complex scientific process which causes thechemical molecules within the vapour mixture to separate according to theirsize. Niamh explains: 'If you can imagine a drainpipe that's ten feet long andyou pour treacle down it, so the inside is coated with treacle, and then youget a box of marbles of different sizes and you pour them down it, the littlemarbles will stick longer than the big marbles. So you get big marbles outfirst and then little marbles. That's what GC does, in a nutshell. Juries canvisualise that, so they go, "Oh, now I get it."'

If the tests show petrol, then, depending on the case, the next step may be tocarry out 'petrol branding'. Most molecules in a can of petrol will evaporateat room temperature (which is why you can smell it), but manufacturers putadditives in their petrol which do not evaporate. The additives make carengines run more efficiently, and can survive very high temperatures. They arealso quite specific to different brands. Additives are extremely stable and canstay on clothes until they are washed out with detergent.

Petrol branding was important in obtaining convictions following one of themost distressing house fires in recent memory. At 4 a.m. on 11 May 2012, a firebegan burning the inside of the front door of 18 Victory Road, Allenton, Derby.Two minutes later, it had raced up the carpeted staircase to the open doorwayof a bedroom full of sleeping children. Their father, Mick Philpott, called 999– 'Help me! My babies are trapped inside the house!' Jade, John, Jack, Jesseand Jayden Philpott, aged between five and ten, died at the scene, and DuwaynePhilpott, aged thirteen, died later in hospital, all from smoke inhalation.

Hours after the flames had been put out, Mat Lee from Derbyshire Fire Servicearrived at the scene. A colleague had found an empty petrol can and a glovenear Victory Road, so Lee was on especially high alert for arson. He removedthe top layer of debris from underneath the front door, and a hydrocarbon dogstarted barking. Lee packaged the material and sent it off to forensic chemistRebecca Jewell for analysis.

Five days after the blaze, the parents of the dead children, Mick and MaireadPhilpott, gave a press conference to thank friends and family for theirsupport. But their behaviour aroused police suspicions. Assistant ChiefConstable Steve Cotterill felt that Mick acted like an 'excited child' insteadof a grief-stricken parent. 'I would have expected him to be completely andutterly destroyed,' Cotterill said later. 'It was a sham, in my view.'

The police put the Philpotts under 24-hour covert surveillance. A bug in thecouple's hotel room picked up Mick telling his wife: 'You make sure you stickto your story,' and later, 'They're not gonna find any evidence, are they? Youknow what I mean?' On 29 May, the Philpotts were arrested on a murder charge(which was later downgraded to manslaughter).

Over a period of six months Rebecca Jewell received various samples from thescene and from the defendants' clothes. In the abandoned plastic tank, shefound a mixture of additives including those from Shell petrol. She foundtraces of petrol in the carpet under the door of the house, but couldn't tellwhich brand it was because the additives were contaminated by a chemical fromthe carpet underlay. She found Shell additive on Mick's boxer shorts and righttrainer. She found Total additive on leggings, a thong and a sandal belongingto Mairead, and on the clothing of Paul Mosley, who had been charged withhelping the Philpotts set the fire.

When the trial began in February 2013 the jury was told that the Philpotts andMosley had started the fire in a bid to incriminate Lisa Willis, MickPhilpott's former mistress. Lisa had spent ten years living in the house withMick, their four children, her fifth child from a previous relationship, andMairead and her children, but had recently left the house and taken herchildren with her to live with her sister. A custody hearing had been scheduledfor the morning after the fire, and Mick Philpott had hoped to pin the arson onLisa, to prevent her winning the right to keep their children. Mick and Maireadhad put all the children to bed in one bedroom, and rested a ladder up againstthe bedroom window. The plan was for Mick to climb up and rescue them, makinghim look like both a victim and a hero. But the fire spread too quickly. Therewas no time to get in through the window and save the children. All three defendantswere found guilty of manslaughter; Mairead and Mosley were sentenced toseventeen years in prison, and Mick to life. The Philpott fire dominated themedia for weeks; the Daily Mail headlined an article 'Mick Philpott: VileProduct of Welfare UK'. While some were wondering if the Philpotts had beenusing the kids to generate their £13 per week child benefits, Niamh Nic Daéid'sthoughts were somewhere entirely different. 'Why didn't the smoke detectorswake the kids?'

One of her Masters students had been part of the team investigating the fire.Together they decided that for his dissertation he would look into the abilityof smoke alarms to wake children. They asked the parents of thirty children toset off the smoke alarms in their properties at random hours of the night.'Eighty per cent of these children did not wake up, even though some of themhad the alarm in their bedroom.' The variable frequency detectors designed toaddress the problem of heavily sleeping children seldom worked. Some of themost effective alarms are reported to be the ones that allowed the mother torecord a message herself: 'So she says, "Get up!" and children respond to thepitch and frequency of her voice.' Now the challenge is to learn the lesson ofthe fire investigators' research – a challenge Niamh's research team is takingup with smoke detector manufacturers.

The desire for custody of children is possibly a unique instigation for arson.Much more common motivations include revenge, insurance fraud or the desire tocover up after a burglary, or even a murder. But people who try to dispose of abody by setting fire to a house or, like Jane Longhurst's murderer, torchingthe body itself (see p.212), are unlikely to succeed. Any forensic investigatordealing with a fire quickly learns to distinguish between the normal effects offire on a body and evidence which may have a more sinister explanation. Whetheror not someone was still alive when the fire started, heat causes the musclesof the body to seize up, drawing the legs and arms up into a classic 'pugilist'stance. Water loss shortens the limbs and causes the body to lose up to 60 percent of its weight. The facial muscles are distorted and the skin of the limbsand torso bursts, creating tears which an inexperienced investigator couldmistake for wounds received prior to death. The bones, made brittle by exposureto heat, often fracture when the body is moved from the scene to the morgue.But, even if a body is badly charred on the outside, it will usually besurprisingly well preserved internally. At a crematorium bodies are reduced toashes by exposure to 815°C heat for around two hours. While structural firescan reach 1,100°C, they generally don't stay hot enough for long enough tocompletely destroy evidence of foul play.

Some people love fires so much that they start them with no obvious motive atall. The pure arsonist. Their addiction starts small but invariably escalates,and is rarely overcome. It often incorporates a sexual element and can befiercely addictive.

One extraordinary serial arsonist started firing Californian buildings in 1984and didn't stop until he was arrested in 1991. During those seven years federalagents estimated he set more than 2,000 fires. Joseph Wambaugh wrote a bookabout him, Fire Lover (2002), and HBO made a feature film, Point of Origin(2002).

The story begins in 1987 when Captain Marvin Casey of the Bakersfield FireDepartment was summoned to a fire in a fabric store. As soon as he got there hewas called to another Bakersfield fire, this time in an arts and craft store.This second one had been extinguished before it overwhelmed the building, andCasey was able to recover a time-delay incendiary device – a lit cigaretteplaced alongside three matches, rolled up in a yellow sheet of notepaper, andheld together with an elastic band. The arsonist had moved the cigarette up soits base was in contact with the match heads, giving him up to fifteen minutesbefore the cigarette burned down, and fire erupted.

Over the next few hours Casey heard of two more fires in Fresno, 100 miles downHighway 99 from Bakersfield. It felt like too much of a coincidence; Caseysuspected that a serial arsonist was in play. Curiously, Fresno had beenhosting an arson investigators' conference which ended shortly before the firesbroke out.

Casey sent the incendiary device from the Bakersfield craft store to afingerprint examiner, who managed to lift a good left ring finger off theyellow notepaper. He put the print through both the state and national criminalrecord databases, but found no matches.

Casey began to think the unthinkable. Could one of the fire investigators atthe conference have set the fires on their way back home? He found out that ofthe 242 officers in attendance, fifty-five had left the conference alone anddriven south along Highway 99. He decided to ask the FBI for help and rangSpecial Agent Chuck Galyan in Fresno. 'Fifty-five names of respected arsoninvestigators? I thought Marv Casey was out in left field somewhere,' Galyansaid. The case went cold. Two years later, in 1989, there was another arsoninvestigators conference in Pacific Grove, followed by another almostsimultaneous outbreak of fires, this time along Highway 101, which hugs thecoast from Los Angeles to San Francisco. Casey couldn't believe it. He workedout that only ten officers with southerly routes home had been at both theFresno and Pacific Grove conferences. This time Chuck Galyan agreed to ask afingerprint expert to check the relevant prints from the state database ofpublic safety professionals. The veteran expert made his comparisons. But hefailed to find a match.

Between October 1990 and March 1991 a rash of new fires broke out aroundGreater Los Angeles, in chain retail stores like Thrifty Drug Stores and Builders'Emporium. Glen Lucero, of the Los Angeles City Fire Department, said, 'Thefires were occurring predominately during business hours. Most arson fires areset under the cover of darkness. This was highly unusual [and showed] a certainamount of bravado and confidence by the person setting the fires.'

In late March, the fires reached their apogee. Five stores were hit on a singleday. The employees of one medium-sized craft store put out the blaze before itproperly caught hold. Investigators found an incendiary device there, still ingood condition, identical to the one Casey had found in Bakersfield four yearspreviously. Six more of these devices were later recovered, a number of them inpillows, giving rise to the arsonist's nickname – the 'Pillow Pyro'.

The investigators knew they were after a clever, experienced and exceedinglydangerous man. He knew enough to start his store fires in the perfect place toencourage their rapid spread. People in these stores were in grave danger ofmeeting the same fate as those who had been trapped in Ole Hardware Store inSouth Pasadena in 1984. That explosive fire had started in amongst polyurethaneproducts, resulting in an inferno that burnt with a blue flame and an eeriehissing sound. Badly burnt bodies were blown out of the building by a flashover– when temperatures reach more than 500°C and all the combustible material inan enclosed space ignites explosively. Four people were killed, including amiddle-aged woman and her two-yearold grandchild.

In April 1991, a 20-strong 'Pillow Pyro Task Force' was set up to liaise withpolice departments across California and track down their man. Threeinvestigators visited Marvin Casey in Bakersfield, who eagerly showed them hisphoto of the fingerprint he had lifted in 1987. Because the print had alreadybeen cleared by an expert, the investigators had low expectations. But thePillow Pyro might have committed a crime in the last four years, so they sentit to Ron George at the L.A. Sheriff's Department.

The Sheriff's Department database had a large collection of fingerprints, ofcriminals, of all police officers in the county, and of anybody who had everapplied for a police job. This time the examiner satisfied himself that he hada match – Captain John Orr, an arson investigator with twenty years' experienceat the Glendale Fire Department behind him. Initially, the investigatorscouldn't believe that he was guilty, and clung to the idea that the fingerprintmust have come from some sort of cross-contamination. On 17 April Ron Georgerang the Pillow Pyro Taskforce and told an agent, 'It's John Orr's. He shouldaknown better. Tell that dummy not to handle the evidence.'

Orr's prints had been on the Sheriff's Department database since he was vettedfor a job as a police officer with the LAPD in 1971. They had been happy withhis prints, but not with a reference from his previous job which had describedhim as 'know-it-all, irresponsible and immature'. Further psychological testsconfirmed his unsuitability for the role and they rejected him unceremoniously.Nevertheless John Orr's subsequent career in the Fire Service had beendistinguished: he had personally instructed more than 1,200 firefighters,organised seminars on fire investigation and written a number of articles forthe American Fire Journal. But how would John Orr have come into contact withevidence at a fire scene in Bakersfield, 100 miles from his base in Glendale?

There was only one unpalatable answer. The task force began surveilling Orr andtalking quietly to some of his colleagues. One of them had been suspicious forsome time. He had noticed that Orr had an uncanny ability to arrive at a firescene before anyone else, and rapidly home in on its origin. (As Niamh NicDaéid explained earlier in this chapter, investigators work scenes inmethodical phases, before approaching the business end.) But most of Orr'scolleagues were incredulous. True, he could be smug when he talked about hisinvestigations, but he was a damn good investigator, and one of their own.

Another conference was soon to be held at San Luis Obispo. The task forcethought Orr might strike and wanted to catch him in the act. Agents watched himall weekend, every hour of the day, but he started no fires. It seemed he couldfeel their eyes on him.

In the end it was Orr's vanity that led to his downfall. He wrote a novel andsent it to a publisher with an astonishing cover letter. 'My novel, Points ofOrigin, is a fact-based work that follows the pattern of an actual arsonist whohas been setting serial fires in California over the past eight years. He hasnot been identified or apprehended and probably will not be in the near future.As in the real case, the arsonist in my novel is a firefighter.' Wheninvestigators got their hands on it, they couldn't believe what they werereading.

The arson attacks mentioned in the manuscript matched many of the Pillow Pyrofires right down to the smallest detail, except for the names. The hero is afire investigator on the hunt for the serial arsonist, Aaron. He compares thetimings of all the fires to the working hours of the firemen in the force andrealises that only Aaron could have done it.

On the morning of 4 December 1991 agents arrived at John Orr's home. Under thefloor mat behind the driver's seat of his car, they found a pad of yellow linednotepaper. In a black canvas bag they found a pack of unfiltered Camelcigarettes, two books of matches, some rubber bands and a lighter.

The day after Orr's arrest, Mike Matassa of the task force made various callsto people he had worked with over the year. One was to Jim Allen, arsoninvestigator and personal friend, who told Matassa, 'You ought to look at theOle's fire. Y'know the one at Ole's Home Center in South Pasadena, October1984? John's obsessed with that one. He was mad when they called it anaccident.' When he got off the phone, Matassa had a flash of recall. Along witheveryone else in the task force he had been reading a photocopy of Points ofOrigin. He remembered that in chapter 6 there was an account of a fire in'Cal's Hardware Store', where five people had died, including a young boy. WhenAaron doesn't get 'credit' for starting it he sets another fire in Styrofoam ina nearby hardware store, to expose the investigators' ignorance. The parallelswere eerie.

On its own, Points of Origin would not have been enough to secure a conviction.But in conjunction with the other evidence – the fingerprint, and a trackingdevice that was installed behind the dashboard of his car – John Orr was foundguilty of twenty-nine counts of arson and four counts of murder. He wassentenced to prison for life with no possibility of parole. He has neveradmitted to any of his crimes. But a fire investigator in Points of Originmakes a telling comment: 'The serial thing usually starts way after they haveexperimented with fire when they're young, and they just continue it if theyaren't caught early. As they grow up, it takes on a sexual atmosphere. Youknow, they are too insecure to relate to people in a direct, person-to-personway and the fire becomes their friend, mentor and sometimes their lover.Actually a sexual thing.'