Seeley International Pty Ltd v Jeffrey

SUPREME COURT OF VICTORIA

COURT OF APPEAL

S APCI 2012 0178
SEELEY INTERNATIONAL PTY LTD (ACN 054 687 035)	Appellant
v
NOEL VIVIAN JEFFREY and HELEN MARY JEFFREY	Respondents

---

JUDGES	WARREN CJ, NETTLE and WHELAN JJA
WHERE HELD	MELBOURNE
DATE OF HEARING	23 September 2013
DATE OF JUDGMENT	18 October 2013
MEDIUM NEUTRAL CITATION	[2013] VSCA 288
JUDGMENT APPEALED FROM	Jeffrey v Seeley International Pty Limited [2012] VCC 1043 (Judge Macnamara)

---

NEGLIGENCE – Product liability – Causation – House fire damaged respondents’ home in 2003 – Competing expert evidence as to the whether the fire was caused by evaporative air cooler or downlights – Whether evaporative air cooler manufactured by the appellant was the cause of a house fire – Whether judge erred in rejecting alternative theories as to cause of fire – Whether findings were against the evidence and weight of evidence – Transport Industries Insurance Co Ltd v Longmuir [1997] 1 VR 125 applied; March v Stramare Pty Ltd (1991) 171 CLR 506; Nom v Director of Public Prosecutions [2012] VSCA 198 referred to – Appeal Dismissed.

---

Appearances:	Counsel	Solicitors
For the Appellant	Mr P J Riordan SC with Mr N D Hopkins	Norton Rose Fulbright
For the Respondents	Mr A J Kelly SC with Mr A D Clements SC	Ligeti Partners

WARREN CJ

NETTLE JA
WHELAN JA:

This is an appeal from a judgment given in the County Court. The judge held that the appellant (‘Seeley’) was guilty of negligence in the manufacture of an evaporative air cooler[1] by equipping the unit with a P0 start capacitor rather than a P2 start capacitor, and that as a consequence the cooler caught fire and caused damage by fire to the respondents’ (‘the Jeffreys’) home. The judge also held that the use of a P0 capacitor rather than a P2 capacitor contravened s 75G of the Trade Practices Act 1974 (Cth). His Honour awarded the Jeffreys damages of $308,000 with interest of $232,689.76 and costs.
[1]In the product warranty, called a Breezair Evaporative Air Conditioner.

The facts

In brief, the facts were as follows:

1) The Jeffreys lived in a house built in 1996. One of the features of the house was a Seeley Breezair Evaporative Air Cooler mounted on the iron roof above the kitchen/meals area.

2) On 21 January 2003, the Jeffreys were at home on annual leave. At about 4.00 am, Mrs Jeffrey was woken by her daughter, Michelle, who said that there was smoke in the house. There had been bushfires the previous day and Mrs Jeffrey believed that the smoke was bushfire smoke coming from outside the house. Accordingly, she turned off the air cooler, which had been running throughout the night, and closed all the windows. Then she went back to bed.

3) When Mr and Mrs Jeffrey woke again at about 8.00 am, Mrs Jeffrey went from the master bedroom to the kitchen to make coffee for Mr Jeffrey. In the process, she noticed a ‘haze’ of smoke in the house. When she brought the coffee up to Mr Jeffrey she observed to him that the house ‘was a bit smokey’. Thinking that the smoke was due to bushfires, Mr Jeffrey told Mrs Jeffrey to turn on the air cooler using the fan rather than the cooling function. He intended that the fan should expel or disperse the smoke. Mr Jeffrey heard the motor in the cooler ‘click on’ (as is recorded in the transcript) or ‘kick on’ (as the judge and counsel heard it). Mrs Jeffrey also opened the house doors.

4) After a very short time, Mrs Jeffrey noticed that the smoke in the house had become thicker. She also noticed ‘a bit of an electrical burning smell, not like wood-fired smell’. She said that she told Mr Jeffrey about it and turned off the air cooler. Then, the television in the master bedroom went off. Mrs Jeffrey went outside because she ‘knew something was terribly wrong’. She could see that the air cooler ‘above the sliding doors ... from the kitchen ... was billowing smoke’. She rushed in and roused Mr Jeffrey from his bed.

5) When Mr Jeffrey emerged moments later, he saw a man whom he did not know pointing a hose up at the roof. The water was not reaching the seat of the fire so Mr Jeffrey grabbed the hose and climbed onto the roof. He was barefoot and clad only ‘in my jocks’ although he was able to walk on the roof. But the hose had ‘shocking pressure’, due to what he believed was a kink in it, and then he slipped off the roof and fell into the rockery.

6) Minutes later, the Country Fire Authority (‘CFA’) attended with multiple units. The first unit arrived at 8:28 am. At or around the same time the motor from the air cooler crashed down through the plaster ceiling, hitting the kitchen table and landing on the floor.

7) The firemen removed part of the roof to introduce hoses into the roof cavity and eventually extinguished the fire. One fireman took a number of photographs of the scene.

8) According to Mr Jeffrey’s recollection, later in the afternoon the kitchen ceiling collapsed. Other witnesses, including a number of experts, believed that the ceiling was ‘pulled down’ by firemen as part of the process of getting to the seat of the fire.

9) Some of the photographs taken at the scene showed that only a relatively small portion of the kitchen ceiling had collapsed and that firemen were gaining access to the roof cavity by a ladder through the hole so created. Another photograph depicted a fireman with a hook on the end of a long handle of a kind used by fire fighters to pull down ceilings. A further photograph showed puncture marks in the plaster ceiling in the vicinity of several downlights mounted in the intact portion of the ceiling. A ceiling may be punctured in that manner to get at the seat of the fire or to test whether the fire is burning above the punctured section of the ceiling.

The witnesses

In addition to lay witnesses, two CFA officers and several experts gave evidence. Captain Wells of the CFA was involved in a supervisory role and had no independent recollection of the matter.

Mr Ranzenhofer was a salaried fire investigator with the CFA. He conducted an on-scene investigation of the fire on the day of the fire. He was at the scene for some two hours and twenty minutes and spent 80 minutes examining within the roof cavity. He found it safe to walk on the beams despite extensive charring. He concluded that the fire originated from the air cooler. He observed that the damage in the roof cavity ‘lessened as it moved away from the unit’.[2] He said that the roof trusses in the vicinity of the unit had burnt away completely.[3] He noted ‘metal discolouration on shroud of unit’, although it later emerged that the discoloured metal object to which he was referring was a separate unit used for heating which was unconnected with the air cooler.
[2]T162, L28-9.
[3]T163.

Mr Ranzenhofer also said that, although he ‘found no evidence to point me in [the] direction of [the downlights]’ he nonetheless ‘investigated [them] as a possible cause of the fire’. He moved some of the Insul-fluff insulation away from them and found no guards around them. But he saw no discolouration in the vicinity of any of them either. That was consistent with the photographs. Mr Ranzenhofer was aware that downlights operate at ‘extremely high temperatures’ but he said that he had never seen evidence of a fire caused by downlights ‘without damage to the external part of it’ and there was none on this occasion. He noted photographs showing dark lines on the un-collapsed portion of the ceiling in the kitchen/meals area, but he said that he did not consider them to be indicative of combustion immediately above. In his view, they were the joins in the plasterboard ceiling which had been covered with tape and then plastered over during installation. They represented a point through which water that had been poured on the fire had tended to seep. He referred to photographs which showed ‘ceiling hook damage’ in the vicinity of one or more downlights in the un-collapsed portion of the ceiling. He said that he had examined the possibility that one of the downlight transformers was the source of the fire. Based on his examination, he had excluded that as a possibility. He said that he looked through the plaster which had fallen from the corner of the room where the ceiling collapsed and found no damage to transformers. He added that, if a transformer were the source of the fire, he would have expected some part of it to have survived and that he could not recall observing the remains of any downlight transformer.

In the CFA Fire Investigation Summary Report which he signed on 29 January 2003, he stated inter alia:

Although not able to categorically ascertain the exact location of the point of origin due to the large amount of damage, it is my belief that the fire initiated in or around the base of the [Breeze Evaporative cooler] unit.

Mr R[4] was a forensic scientist specialising in metallurgical failure, investigation, materials, explosion and fire related investigations. He attended the scene, probably the day after the fire, at the request of the Jeffreys’ household insurer. He recovered the electric motor from the air cooler unit and what he described as ‘electrical system remains’ for further examination. In his preliminary report to the insurer dated 28 January 2003 he concluded that:

The fire had initiated in the roof space above the main living area adjacent to the kitchen and was directly associated with the evaporative cooling unit.

Mr R also noted severe damage to the unit and remarked on ‘artefacts of localised overheating associated with the rotor shaft bearings’.

[4]So called because of the need to conceal his identity.

In a further, lengthier report to the insurer dated 31 October 2003, Mr R reiterated his initial conclusion that the fire was caused by ‘bearing overheating in service’ and added that:

Further, fusion of conductors ultimately arising from high resistance current tracking across the gradient insulation on wiring within the unit cannot be excluded as a contributing ignition source. All other causes of the fire have been excluded.

Like Mr Ranzenhofer, Mr R excluded the downlights as a possible cause of the fire. He said that he found all five of them and examined them; although, as the judge observed, the fact of that examination was not recorded in Mr R’s report. Mr R explained, however, that it was his practice to record only those things which he regarded as significant, not those which he excluded as being causally insignificant.

Mr Moore was a metallurgical expert who held the Degree of Bachelor of Applied Science in Metallurgy and a Graduate Diploma in Engineering and Building Fire Safety and Risk Engineering. He was a specialist fire examiner. He was engaged in June 2004 by the Jeffreys’ solicitors to advise on Mr R’s opinion and as to causal issues relative to the fire generally.

Mr Moore rejected Mr R’s view that the bearings initiated the fire. He said that the damage to the bearings was: ‘considered to be a localised change in the colour of the oxidation and blueing possible [sic] due to molten plastic having run down the side of the bearing’, and that the shaft associated with the relevant bearing ‘did not exhibit any signs of significant heat damage or wear due to rough or out of balance running of the bearing’.

Mr Moore identified a damaged metal strip which showed signs of electrical arcing as being the cause of the fire. At that stage, he believed that the strip was made of copper. Later, after it had been cleaned, he discovered that it was made of mild steel. On that basis, he formed the view that it had come from the motor capacitor in the air cooler. Thus, in a report dated 27 December 2007, Mr Moore opined:

Given the level of metal loss it is considered that the damage to the strip would have occurred over a period of time and multiple arc events but did not trip the circuit breaker. Given the strip comes from within a control box it is considered more likely that an external fire would cut the power to the unit prior to affecting the strip that is buried within the control box and hence surrounded by plastic that would melt around the strip, also the fire would be more likely to affect the conductors within the unit that would be more exposed.

Dr Peter Hart was an electrical engineer retained by the Jeffreys’ solicitors. In his first report to the Jeffreys’ solicitors dated 20 August 2008, he reviewed the existing expert opinions and concurred in the rejection by Mr Moore of the idea that the fire was caused by an overheated bearing. Based on the evidence then available to him, Dr Hart postulated five possible causal scenarios:

1) The first entailed overheating in the triac circuit causing the plastic box in which the triac was located to overheat and drape itself onto the circuit board resting on the heat sink or the toroidal transformer. That was said to be consistent with the smell and the smoke which was discovered at 4.00 am.

2) The second was that the toroidal inductor 1NI overheated in its housing and caught fire. The hypothesis was that: ‘The plastic enclosure may have softened and draped over the circuit’.

3) The third was that the transformer TF1 overheated and caught fire. Once again, the idea was that the plastic enclosure softened and draped over the circuit.

4) The fourth entailed an electrical fault developing in the ceiling space resulting in an electrical short circuit.

5) The fifth was that the air cooler motor start capacitor failed and caught fire. The hypothesis was that the failure resulted from ‘gradual deterioration due to internal heating’. Dr Hart considered that the heating could have caused a ‘localised failure of the polypropylene separator which led to the build up of organic gases and eventually the capacitor ruptured. This was the cause of the smoke haze through the house’. Dr Hart said that:

When the motor was turned on again at 8:00 am ... the capacitor was re-energised and generated further smoke before catching fire. The internal polypropylene layer in the capacitor is combustible.
…
The capacitor had ruptured but was operational to a limited degree.
…
Vaporised regions may not draw a significant fault current because of the fineness of the conductive layer and hence the high electrical resistance that limits fault current level. It is possible but unlikely that the motor operated without apparent problems and blew air normally. I expect some degradation of motor performance would have occurred however, this may not have been noticeable to the occupants.

Dr Hart concluded that, of the five possible scenarios, one and five were the most likely, because:

The apparent absence of any motor capacitor parts (such as the terminals or the metal can) in the debris is consistent with the thesis that the capacitor exploded and caught fire.

At that stage, he also considered that scenario one was more likely than scenario five, because:
The air cooler had been on for an extended period because the ambient temperature conditions had been very hot. ... The heating potential of the triac and toroidal inductor would be greatest at some intermediate speed setting. Unfortunately there is no report about the air flow setting.

Following very late discovery by Seeley that the ICAR ecofill capacitor model MLR25PRL used in the air cooler contained parts made of steel, Dr Hart wrote a further report dated 19 April 2012 in which he concluded on the basis of that newly discovered information that he was now almost certain that the cause of the fire was Scenario 5:

I am [now] almost certain that the part Mr Moore inspected is from the motor capacitor of the Jeffreys’ evaporative cooler.

The damage to the strip was caused by a ‘dancing arc’:
…due to clashing between two electrically live metal parts; that is short circuit. The result is localised melting erosion of metal due to intense heating.

The possibility of liquid metal attack could safely be excluded on that the basis that:
There is no evidence that there was molten aluminium in the vicinity of the capacitor. The housing did not melt at that end. The strip was protected by the plastic capacitor body.

And, because the central strip was contained with a plastic insulating layer, the failure mechanism would have been the result of micro-breakdown in the insulating polypropylene casing:
… the contact must have been from inside. Because the central strip is normally inside a plastic insulating cylinder, there must have been an explosive failure of the part for this damage to occur. The short circuit probably involved direct contact between the central strip and the other terminal metal.
…

The failure mechanism of polypropylene film capacitors is that micro-breakdowns of the insulating film occur and this liberates some organic gas. A pressure build-up can occur but eventually leads to rupture. The gas is flammable because it is derived from the plastic film. Because the capacitor has live voltage across it, a spark is certain to occur. The gas ignites and the capacitor explodes and burns.

Seeley called two expert witnesses, Mr Kutek, who held qualifications in metallurgy and engineering, and Mr Baghurst, who was an electrical engineer and adjunct associate professor in the University of New South Wales.

To begin with, Mr Kutek was particularly critical of Mr R’s rejection of the possibility that the fire was caused by downlights igniting Insul-fluff insulation and critical of Dr Hart’s conclusion that the metal strip came from the air cooler. He described Dr Hart’s five scenarios as ‘essentially unproven and unprovable hypotheses made by a person who lacked the fundamental knowledge of fire behaviour and interpretation of evidence from a fire-scene to evaluate it with his own hypotheses’. In a report dated 25 March 2009, Mr Kutek claimed that:

having examined samples of components that were used in the manufacture of the subject air cooler, I can say that the strip was not part of the cooler as supplied.

Later, however, in the course of cross-examination, Mr Kutek ultimately conceded that the metal strip probably had come from the motor start capacitor in the air cooler.[5]
[5]T.1462.6-.20.

Mr Baghurst was also critical of Dr Hart’s analysis. He rejected the idea that it was possible that the fire was caused by the triac or the toroid, because he said they were protected, and he rejected any possibility of the fire having been caused by the capacitor because, at that stage, he said that he thought that the capacitor was of P2 type (which is designed to fail safely) and not P0 type (which is not so designed). In his view, the most likely cause of the fire was the downlights overheating and igniting the Insul-fluff insulation in the roof cavity.

Subsequently, when it was discovered that a P0 type capacitor had been used in the construction of the air cooler, and that was brought to Mr Baghurst’s attention, Mr Baghurst accepted that the metal strip was likely to have come from that capacitor. At that point, however, he developed a different set of reasons to support his contention that the fire was not caused by the metal strip:

Note that the steel strip found in the ICAR capacitor [which is a P0 capacity] is located at the centre, and runs along the longitudinal axis of the capacitor (which has a cylindrical shape)...it is embedded at the centre of the capacitor in a heavy resin coating, which is, in turn, covered with a plastic sleeve. In order for electrical arcing to take place, another solid metallic ‘electrode’, electrically connected into the motor circuitry, would be required. No such electrode exists in a capacitor of this type, the nearest electrically conductive material to that central metal strip being the capacitor elements themselves...

Mr Baghurst also claimed that he had never experienced any capacitor failure ‘which remotely resembles’ the kind of dancing arc mechanism which Dr Hart described, and for good measure he characterised Dr Hart’s theories as ‘an ever increasing frenzy of baseless speculation’.

The judge’s reasoning
(i) The Downlights

The judge rejected Seeley’s case that the fire was caused by downlights igniting roof cavity insulation. As his Honour observed, there were a number of considerations which rendered that improbable. They included that there was no explanation of why smouldering should first occur in 2003 when the Insul-fluff insulation had been in place since 1998. As was earlier noted, the CFA photographs (which appeared to have been taken at about 10.10 am on the day of the fire) showed a fireman using a hook to pierce the ceiling adjacent to four of the five downlights to check whether there was any combustion near the lights, and it appeared that there was not.

Seeley made much of the fact that there was no photograph of the fifth downlight, or at least that it was not clear that what was identified in one of the CFA photographs as the fifth downlight was what it purported to be. There was, however, no evidence suggesting initiation by the fifth downlight, and both Mr Ranzenhofer and Mr R swore that they had examined the downlights and excluded them as a possible source of ignition. Admittedly, as the judge observed, they were to some extent reliant on their memories of what they observed because they had not made contemporaneous notes of their downlight examinations. But each of them had filed written reports shortly after their examinations of the premises and the contents of those reports were at least consistent with them having excluded the downlights as a possible cause of the fire. As has been noted, Mr R filed his report on 21 January 2003, which he confirmed in further reports filed on 28 January 2003 and 31 October 2003. Mr Ranzenhofer filed his report on 29 January 2003. Thus, as the judge observed, it was plausible that they had both examined and excluded the downlights as a possible source of ignition and, in the case of Mr Ranzenhofer, the probability of that being so was enhanced by his complete disinterest in the outcome of the case.

The judge also noted that Mr Jeffrey had been able to walk on the roof barefooted as he attempted to put out the fire and, in the judge’s view, that fact told against the flames having flared up outside the footprint of the unit where they might have heated up the surface of the steel roof sheeting.

(ii) The toroid, transformer and triad

The judge found that it was equally improbable that the fire was caused by the toroid, the transformer or the triad. Apart from anything else, as his Honour observed, the theory that any of those components was the cause of the fire was premised on a false assumption that the air cooler was labouring overnight in the face of high ambient temperatures. The evidence was that a cool change had arrived on the evening of 20 January 2003.

(iii) The start capacitor

The judge was persuaded, however, that it was more likely than not that the fire was caused by the start capacitor. Four factors combined to lead his Honour to that conclusion. First, despite Mr Kutek’s initial asseverations of confidence that the metal strip could not possibly have come from the capacitor, ultimately he had to accept that it probably had come from the capacitor. As the judge said:

Given his concession that the strip ‘probably’ came from the motor capacitor in the Seeley unit, with this view of things supported ultimately by Mr Moore and Dr Hart, I believe I should accept this rare concordance of expert opinion.

Secondly, while the experts were agreed that the failure of a capacitor in short circuit was an unlikely and atypical event, international standards which graded motor capacitors according to their propensity to fail safely (scil. in open circuit), or unsafely (scil. in closed circuit), demonstrated that failures of capacitors in closed circuit do occur.

Thirdly, there was evidence that a senior technical officer of Seeley had testified in a previous proceeding concerning the alleged failure of another Seeley unit, that:

The statements that ‘capacitors burn readily’ needs clarification. Capacitors are made from fire retardant polymers. These polymers will not support fire without an external energy source. This energy source could come from arcing between foil layers within the capacitor, which would progressively consume the capacitor by driving off the fire retardants and then allowing the polymer to oxidise. The capacitor is unlikely to be visibly on fire at this stage. However, if the heat and energy of the arcing could get to polymers not so protected, fire could readily propagate. The problem with most fire retardants is that they are volatile and are driven off by the fire. Once they are driven off, the polymer is left unprotected and is then consumed. Hence there is no trace left of the cap or electrical boxes, or the capacitor itself. All these components are made of good fire retardant plastics, but are all consumed by the fire once the retardants are driven off. About all you can conclude is that since they are completely gone, and lesser polymers in other parts remain, then these parts were probably near the seat of the fire.

As the judge said, that was an acknowledgment that motor capacitors do have the ability to create fires.

Fourthly, it was Mr Baghurst’s opinion that when capacitors rupture they emit a very audible pop, bang or explosion. Since the Jeffreys were both awake at 8.00 am on the morning of the fire when the unit was switched back on, it was unlikely to have popped then. It was more likely that it occurred at or about 4.00 am when the Jeffreys’ daughter woke and informed her mother that it was ‘smokey’ in the house. It may have been the pop which woke her although she was not conscious of it.

Fifthly, despite Mr Baghurst’s evidence that he found it ‘quite inconceivable’ that a failed capacitor once reactivated could result in the sort of dancing arc which Dr Hart envisaged, or how the capacitor could have restarted the motor at 8.00 am after failing at 4.00 am, he did not explain why he found it ‘quite inconceivable’ and counsel for Seeley was unable in final address to say why it should be regarded as inconceivable.

Finally, the judge reasoned that, if one put together what his Honour described as both of Mr Baghurst’s hypotheses, namely, that the metal strip probably came from the start capacitor and that it probably suffered arc damage, with Mr Moore’s evidence that there was no obvious source of molten metal in the near vicinity which would explain arcing, the conclusion that the capacitor was the cause of the fire was difficult to escape.

Grounds of appeal

There are four grounds of appeal:

1) The judge erred in finding that the start capacitor caused the fire.

2) The judge erred in failing to find that it was more probable than not that the fire was caused by downlights fitted in the ceiling of the Jeffreys’ home.

3) The judge engaged in an incorrect process of reasoning by rejecting alternative theories of the cause of the fire.

4) The judge’s findings of fact were against the evidence and the weight of the evidence.

Ground 1 – Whether fire caused by capacitor
(i) Mr Baghurst’s opinion

In support of Ground 1, Counsel for Seeley contended that the judge erred in finding that the capacitor caused the fire by overlooking evidence given by Mr Baghurst that the strip could not have suffered arc damage; and by proceeding, contrary to that evidence, on the basis of a false premise that Mr Baghurst was of the view that the strip had suffered arc damage and had come from the capacitor.

We do not accept that submission. Mr Baghurst was not always of the view that the metal strip could not have suffered arc damage. On the contrary, to begin with he agreed with Mr Moore that damage to the strip was consistent with arc damage and that the consequent heat had ignited nearby combustible material.

In his report of 13 May 2005, Mr Moore wrote:

Examination of the contents of the plastic bag of miscellaneous debris at the offices of Robert Barnes revealed ...
a contact with a copper strip attached. … Cleaning of the copper strip … [revealed that it] was actually made from mild steel and not Copper. …
Mild steel would not be susceptible to liquid metal attack from molten aluminium in the same manner as copper. Mild steel also has a higher melting point.
The areas of damage are very localised and the strip has not suffered any melting over an extended area. …
The damage to the strip is consistent with prolonged arcing that has melted and heated the strip within the vicinity of the arcing. It is considered that this heating then ignited nearby combustible material such as any plastic housing material or dust and lint.[6]

In a response, Mr Baghurst stated in his first report dated 23 February 2009 that he agreed with Mr Moore that the metal strip played an active role in the initiation of the fire:
… I disagree [with Dr Hart] that the ‘iron strip’ may be excluded. My thorough examination of a Breezair unit of the type in question showed that no such strip forms part of the construction of those units, and it is quite possible that it formed part of the construction of the downlighting system.
…
I agree with Hart that it is quite unusual for the steel strip discussed by the other experts and which had an electrical connection, to constitute a component which normally carries electric current. It is quite usual, however, for steel to be used in earthing circuits, and it is possible therefore that the steel strip formed part of the earthing arrangements for a downlight supply transformer….
…
… In summary, I agree with Mr Moore that the metal strip may well have played an active role in the initiation of the fire, but I can say with confidence that the metal strip was not part of the air-cooler unit.[7]

[6]AB Vol 3, C309-10.

[7]AB Vol 5, C 1230.

At that stage, Mr Baghurst dismissed any possibility of the metal strip having come from the capacitor because he said he believed that the capacitor which had been fitted to the air cooler did not contain any ferro-magnetic parts. It was only later, when it was revealed that the air cooler start capacitor did contain ferro‑magnetic parts and, as a result, Mr Baghurst was obliged to concede that the metal strip had come from the capacitor, that he changed his opinion about the metal strip having played an active role in the initiation of the fire. As has been seen, he then claimed that it could not have done so because, as embedded in the capacitor, it would have been surrounded by a heavy resin coating which precluded arcing. Hence, in a letter of 24 February 2012 to Seeley’s solicitors, Mr Baghurst wrote:

I now agree with Dr Hart that the metal strip in question may have come from the motor capacitor.
… As discussed in detail in Section 3, although I have found one particular brand of capacitor which incorporates a metal strip which is in some ways similar to that which was found at the fire scene, and which may possibly have been used in the evaporative cooler in question, the possibility that there was arcing at any such strip is very remote.[8]
…
In the ICAR capacitor, I found a metal strip which was not unlike that which has been described in various [of the] reports on this case. The strip was of steel (confirmed as ferromagnetic), and is shown in Photographs 3 and 4.
…
Note that the steel strip found in the ICAR capacitor is located at the centre, and runs along the longitudinal axis of the capacitor (which has a cylindrical shape). As shown in Photograph 4, it is embedded at the centre of the capacitor in a heavy resin coating, which is, in turn, covered with a plastic sleeve. In order for electrical arcing to take place, another solid metallic ‘electrode’, electrically connected into the motor circuitry, would be required.
No such electrode exists in a capacitor of this type, the nearest electrically conductive material to that central metal strip being the capacitor elements themselves, consisting of a cylinder formed from plastic film material upon which has been deposited microscopically thin ‘metalised’ layers which would be incapable of sustaining arcing of the type which is claimed to have caused erosion of the metal strip in question.

[8]AB Vol 5, C 1281.

There is also no reason to doubt that the judge was mindful of Mr Baghurst’s change of view. In the passage of the judge’s reasons which counsel for Seeley submitted was indicative of error, his Honour said that:

Once one concludes that the steel strip was probably from the capacitor, as conceded by Mr Kutek, the only other explanation as to how it sustained the damage observed was molten metal attack. Mr Moore’s view was that there was no obvious source of molten aluminium in the near vicinity which would explain the damage. Therefore, in the regrettable circumstances flowing from the loss of the strip itself, it is easy enough to accept the view which commended itself to Professor Baghurst, when he believed that the strip probably emanated from a downlight transformer, that it exhibited arc damage. If one puts both of Professor Baghurst’s hypotheses together, viz that the strip probably came from the capacitor and that it probably suffered arc damage, the conclusion that it was the cause of the fire is difficult to escape.[9]

[9]Reasons, [156] (emphasis added).

As we read that section of his Honour’s reasons, it is apparent from the emphasised section that his Honour well understood that Mr Baghurst had altered his view about the nature of the damage to the strip after it was revealed that the strip had come from the capacitor. It is also apparent from questions which the judge put to Mr Baghurst in the course of his cross-examination, and to which we shall refer again later in these reasons, that Mr Baghurst’s change of mind was an important part of his Honour’s reasoning process.

It is evident, too, from the passage set out above that the judge preferred Mr Baghurst’s initial view to Mr Baghurst’s more recently posited thesis of ferro‑aluminium deposits, because Mr Baghurst’s initial view accorded with Mr Moore’s metallurgical microscopic examination of the strip itself. In contrast, as it emerged in cross-examination, Mr Baghurst’s latter day aluminium deposit thesis was based on no more than his having studied metallurgy as an undergraduate in the 1960s as one of the subjects offered in his electrical engineering degree and having viewed a photograph of the strip after it was revealed that the strip had come from the capacitor.[10]
[10]T.1605.27-1607.6.

It will also be recalled that the Seeley senior technical officer who had given evidence in a previous proceeding concerning a failed Seeley air cooler (whose testimony was tendered in this proceeding) deposed that there can be arcing between foil layers within the capacitor which progressively consumes the capacitor by driving off the fire retardants and allows the polymer to oxidise resulting in the rapid propagation of fire.

In the circumstances, his Honour’s reference to and preference for Mr Baghurst’s hypothesis that the strip ‘probably suffered arc damage’ can only have meant, and been intended to mean, Mr Baghurst’s initial view.

(ii) Probability theory

Counsel for the appellant submitted that, even if that were so, the judge committed a ‘grave error’ by aggregating the probabilities of the two hypotheses – of the plate coming from capacitor and the plate causing arcing – in order to come up with the conclusion that it was more likely than not that the plate caused arcing which resulted in the fire. According to counsel, basic probability theory demanded that, even if the probability of each of the two events under consideration were as high as 70%, the probability of them coinciding was the product of their individual probabilities, which is to say only 49%.

We do not accept that the judge committed any error of the kind alleged. In our view, counsel’s invocation of probability theory is misplaced. The mathematical function on which he relied is one for the determination of the probability of two or more events of individual given probability occurring simultaneously. In contrast, this was a circumstantial case in which the known coincidence of circumstances was prone to give each individual circumstance greater significance than it would have had by itself. His Honour’s observation as to putting Mr Baghurst’s hypotheses together and coming up with a conclusion that the capacitor was the cause of the fire is to be seen in that context. It does not stand alone but comes at the end of a detailed analysis of the evidence as to all of the factors which individually and more so in combination led to the conclusion that the capacitor was causative. On any reasonable view of the matter, the evidence so analysed was clear that several highly significant events had occurred simultaneously. The fact of their coincidence added to, not subtracted from, the probability that in combination they were causative of the calamity.

Counsel for Seeley contended that there were, however, a number of other errors in the judge’s process of reasoning. First, in counsel’s submission, it was not open to find that the capacitor was causative of the fire unless it were first found that the capacitor exploded in the hours before the fire and so failed in short circuit. Yet, counsel said, the evidence was that the occurrence of each of those possibilities was decidedly atypical and improbable. Hence, applying basic probability theory to determine the likelihood of the capacitor causing the fire, it was necessary to multiply together the individual probabilities of each of those two atypical and improbable possibilities, with the result that the only conclusion to which to the judge could properly have come was that the possibility of the capacitor being causative of the fire was highly improbable.

We reject that argument, too. It proceeds upon the same misconception that the exercise for the judge was one of ascertaining the mathematical probability of the coincidence of two random events. That was not the nature of the exercise. As we have said, this was a circumstantial case in which the coincidence of circumstances was capable of conferring greater significance on each individual circumstance than it might have had by itself. More precisely, as Gibbs CJ and Mason J explained in Chamberlain v The Queen [No2],[11] the process was that:

At the end of the trial the jury must consider all the evidence, and in doing so they may find that one piece of evidence resolves their doubts as to another. For example, the jury, considering the evidence of one witness by itself, may doubt whether it is truthful, but other evidence may provide corroboration, and when the jury considers the evidence as a whole they may decide that the witness should be believed. Again, the quality of evidence of identification may be poor, but other evidence may support its correctness; in such a case the jury should not be told to look at the evidence of each witness 'separately in, so to speak, a hermetically sealed compartment'; they should consider the accumulation of the evidence.
…
It follows from what we have said that the jury should decide whether they accept the evidence of a particular fact, not by considering the evidence directly relating to that fact in isolation, but in the light of the whole evidence, and that they can draw an inference of guilt from a combination of facts, none of which viewed alone would support that inference.

[11](1985) 153 CLR 521, 535-6.

Of course, Chamberlain concerned circumstantial reasoning in a criminal context. But it is the same in a civil case except that the standard of proof is less demanding. As Lord Cairns observed in the Belhaven and Stenton Peerage case,[12] the jury, or in this case the judge, must consider the weight to be given to the united force of all the circumstances put together:

[I]n dealing with circumstantial evidence, we have to consider the weight which is to be given to the united force of all the circumstances put together. You may have a ray of light so feeble that by itself it will do little to elucidate a dark corner. But on the other hand, you may have a number of rays, each of them insufficient, but all converging and brought to bear upon the same point, and, when united, producing a body of illumination which will clear away the darkness which you are endeavouring to dispel.

[12]Belhaven and Stenton Peerage (1875) 1 App Cas 278, 279 (Lord Cairns LC).

To similar effect, as Tadgell JA stated in Transport Industries Insurance Co Ltd v Longmuir:[13]

… to assess the evidence in a case like this by reference to various individually-pleaded particulars, as though running through items on a check list, is apt to mislead. The evidence is to be evaluated as a whole in order fairly to consider whether the party bearing the onus of proof has established what is ultimately sought to be proved. The object of the exercise of evaluation is to discover whether the evidence paints a picture reflecting real life, rather than to place a tick or a cross against paragraph after paragraph of torpid pleading. A true picture is to be derived from an accumulation of detail. The overall effect of the detailed picture can sometimes be best appreciated by standing back and viewing it from a distance, making an informed, considered, qualitative appreciation of the whole. The overall effect of the detail is not necessarily the same as the sum total of the individual details.[14]
In a civil case like this, where there is no direct evidence of a fact that a party bearing the onus of proof seeks to prove, ‘it is not possible to attain entire satisfaction as to the true state of affairs’.[15] In such a case, however, the law does not require proof to the ‘entire satisfaction’ of the tribunal of fact. A definition of the sufficiency of circumstantial evidence in a civil case to support proof by inference from the directly proved facts was given by the High Court in the unreported case of Bradshaw v McEwans Pty. Ltd[16] in a passage since repeatedly adopted: e.g. Luxton v Vines;[17] Holloway v McFeeters;[18] Jones v Dunkel;[19] Girlock's case.[20]

[13][1997] 1 VR 125, 141; see also Nguyen v Cosmopolitan Homes [2008] NSWCA 246, [54]-[55].

[14]Cf Hall (Inspector of Taxes) v Lorrimer [1992] 1 WLR 939, 944; Shepherd v R (1990) 170 CLR 573, 579-80.

[15]Girlock (Sales) Pty Ltd v Hurrell (1982) 149 CLR 155, 169 (Mason J).

[16](1951) 217 ALR 1, 5.

[17](1952) 85 CLR 352, 358.

[18](1956) 94 CLR 470, 480-1.

[19](1959) 101 CLR 298, 304.

[20]Girlock (Sales) Pty Ltd v Hurrell (1982) 149 CLR 155, 161 and 168.

Here it was known that several highly significant events had occurred simultaneously. Hence, standing back and viewing the picture from a distance, and making an informed, considered, qualitative appreciation of the whole in the way that Tadgell JA explained in Longmuir, the coincidence of those individual events added to, as opposed to subtracting from, the likelihood that in combination they were causative of the calamity.

Counsel for Seeley relied on the fact that the capacitor would probably have made a noise when it exploded and yet that no member of household was conscious of the noise. Counsel suggested that it would have been a ‘loud’ noise. But there
was no evidence of how loud it would have been; and Dr Hart was careful to resist both the judge’s and counsel’s suggestions that it would have been loud:

HIS HONOUR: If such an explosion occurred supposing I was lying in bed in the house where it happened would I hear a big bang or would there be no bang or what would be the situation? --- There would be some noise from it, it’s inside an enclosure which is – inside the capacitor enclosure, it’s up in the roof, I’m not saying - - -
HIS HONOUR: Or above the roof level as we have been reminded from time to time? --- Yes.[21]
…
HIS HONOUR: If there was an explosion of the capacitor of the sort of energy you’re talking about there would be a loud explosion noise, wouldn’t there ? --- There would be some noise, I agree with that, I can’t tell you that I have been in a place where a capacitor has exploded.
HIS HONOUR: But there would be a loud noise? --- There would be some noise, yes.[22]

[21]T.959.11-.18.

[22]T.965.1-.7.

Furthermore, as the judge remarked, if the capacitor failed at 4.00 am in the morning when all of the household were asleep, they might very well not have heard it; and, as the judge conjectured, presumably on the basis of Dr Hart’s observation in evidence to the same effect,[23] that is one possible explanation for the fact that the Jeffreys’ daughter, Michelle, woke up at that time.
[23]T.965.20-.24.

In any event the ascertainment of causation did not require the judge to be certain or even persuaded on the balance of probabilities as to everything concerning the events in issue before him. In the end, whatever uncertainties or mysteries remained as to the individual facts and circumstances, it was sufficient for his Honour to decide, as he did, that he was satisfied, in the sense of reaching a state of actual persuasion,[24] that the concatenation of circumstances of which there was evidence rendered it more likely than not that the capacitor caused the fire.
[24]Nom v Director of Public Prosecutions [2012] VSCA 198, [106]-[112].

(iii) Dancing arc theory

Counsel for Seeley argued that the judge was at fault in failing properly to address what counsel characterised as the sheer improbability of Dr Hart’s dancing arc theory of conflagration. In counsel’s submission, to accept Dr Hart’s thesis required one to accept the highly implausible suggestion that the start capacitor exploded sufficiently to breach its internal section as well as the two protective layers around the steel strip at the centre of the capacitor, thus exposing the metal strip; another unidentified and exposed wire (or the terminal) ‘slid down’[25] into the newly formed hole to create a dancing arc; the arc continued to dance for some time despite the fact that the copper from the wire would have been beading and shortening, causing a fire which was contained within the control box; the fire went out and stayed out until 8.00 am (despite smoke appearing in the house the next morning); and then the capacitor caught fire again within five minutes of the cooler being restarted.
[25]T.960.3-5.

The difficulty with that argument, however, is that, although counsel for Seeley strove vigorously in the cross-examination of Dr Hart to establish that his thesis was implausible in the several respects suggested, Dr Hart gave sound reasons for rejecting each criticism and, therefore, to conclude that his thesis was plausible. Thus:

Counsel: No, but you’re saying there was an explosion that was sufficient to cause a fire at some stage prior to 4 a.m.? --- Yes, it ruptured – in my view, that ruptured the capacitor and there was a fire.[26]
[26]T.944.15-.18.
…
Counsel: All right, and that would just start up the air conditioner, no problem at all, it wouldn’t be an issue? --- No, the – it will start the motor not the air-conditioner, the motor, and I’m not saying that it would be normal operation but the motor would turn and blow air.[27]
[27]T.945.12-.16.
…
Counsel: So the problem with all these bits you sort of try and link together is that you’ll fall over at the last hurdle, because a short circuit failure resulting from an explosive failure is going to lead to abnormal operation of the air conditioner, noticeably abnormal, and that’s not what happened? --- Well, I don’t – that’s your term. I asked Michelle [Jeffrey} specifically whether the operation was normal, and she didn’t say it was normal. She said it was delivering air.[28]
[28]T.945.22-.30.
…
Counsel: Explain the short-circuit then? --- The failure mechanism is a degradation of the polypropylene film, the plastic film. Over time micro‑breakdowns causing a little bit of gas to be released and towards the end of the capacitor’s life this breakdown can become more frequent so more there’s more gas produced, a build up of gas, eventually rupturing the case and fire and that of course damages the capacitor either a little bit or a lot. And so then I’m saying that there is the possibility that there is a short circuit involving a central electrode which has been exposed and another part of another connection, the other side.[29]
[29]T.956.7-.18.
…
HIS HONOUR: … So the theory of this fire that you have said is the most likely one is that there has been a failure involving a closed circuit? --- Yes.
HIS HONOUR: And this is the point on which you’re being cross-examined, Mr Hopkins, yesterday evening [you] said well, if there was an explosion or an explo[d]ing failure it would be so dramatic you would have had in effect an open failure rather than a closed failure, and the answers you’re giving are in response to that? --- Yes, and my answer is that in most cases I would expect to see an open circuit failure but the possibility of a short-circuit failure is certainly there and it does explain – it’s a plausible scenario in my opinion.
Counsel: But it’s the least likely of the scenarios in terms of the failure mode, isn’t it, in terms of open or short circuit? --- Yes, it’s the least likely but it’s a possible scenario.[30]
[30]T. 957.13-.16.
…
Counsel: Talk me through that, if the wire connected to the other terminal touches the internals, how does that work? --- The capacitor is severely damaged as a result of failure, the wires are in close contact, close location, and they have – one wire is – the terminals have perhaps blown apart, one terminal and its wire are now resting on the central strip.
...

Counsel: How does the wire - - - ? --- The capacitor has blown apart so there are two – in my scenario, the two terminals which have wires connected to them.
Counsel: So four terminals here, so it’s blown off one of these two terminals? --- There are two sets of terminals which are in parallel, they are in pairs. There are two wires connected to each set of terminals, one of the terminals is connected to the central strip, the other terminal is not connected to the central strip it’s just connected at the top. The capacitor’s boom, blown apart, somehow we have had the short terminal resting against the central strip and that’s the short circuit, and that would explain why the motor would run, it explains why there is what I say is an arc - - [31]
[31]T.958.20-.T.959.3.
…
HIS HONOUR: By com[ing] apart you mean the polypropylene has been stripped off by some means exposing these terminals? --- The device has exploded, your Honour, there is a graduation of explosions or failure modes, it may be a little hole, in some cases it may be a complete explosion of the capacitor.
HIS HONOUR: If such an explosion occurred supposing I was lying in bed in the house where it happened would I hear a big bang or would there be no bang or what would be the situation? - - - There would be some noise from it, it’s inside an enclosure which is – inside the capacitor enclosure, it’s up in the roof, I’m not saying - - -
HIS HONOUR: Or above the roof level as we have been reminded from time to time? --- Yes.[32]
[32]T.959.5-.18.
…
Counsel: So this means the short wire on one end you’re saying has somehow slid all the way down and touched the wire? --- Slid is your word, I can’t tell you exactly how it got there, if we had a video of it we would know.
Counsel: Then it dances around just on that spot? --- Touching.
Counsel: It dances around on that spot, it doesn’t dance off, it continues to dance? --- That’s correct because the parts are constrained by the physical situation so there may well be some reason why it’s got to that point, and it’s just sitting there and it’s making a sparking contact. …
Counsel: That happens and then in the morning at 8 o’clock it gets turned on and this dancing wire - - -? --- Well, it’s just dancing when the motor is on, it’s had a rest overnight.[33]
[33]T.960.3-.T.960.25.
…
Counsel: Mrs Jeffrey said she heard the motor kick in, it started up? --- Yes.
Counsel: Not possible, it is? --- It is, in fact this is why I brought the book in because if there is a connection through the capacitor, if it’s shorted the motor will start.
Counsel: Yesterday I said, ‘It couldn’t possibly be normal operation?, and you said, ‘No, no, I agree’? --- That’s correct, the capacitor is there for a reason, it increases the performance of the motor but that‘s not to say the motor won’t run without a capacitor; it will.
Counsel: Obviously you have had a lot of opportunity to think about this overnight, let’s go back and work our way through the theory. When do you say the explosion occurred in this theory? --- During the night.
Counsel: Before 4 a.m., before they turned it off? --- Yes.[34]
[34]T.960.30-.T.961.14.
…
Counsel: Then at 8 o’clock in the morning when they turn it on? --- M’mm.
Counsel: It operates, you say? --- Yes, it starts.[35]
[35]T.961.28-.30.
…
Counsel: How does the fire start then? --- Because the circuit’s engaged again, it starts to spark again and there is some combustible material that is set on fire.
Counsel: What’s that? --- It’s in a plastic environment.[36]
[36]T.962.4-.7.
…
How can you run the theory? --- Capacitors explode, I know that.
Counsel: That’s not answering my question, how can you run a theory about the plastic box catching fire if you have never even looked inside it or seen it? --- Because I know that polypropylene motor capacitors are prone to fail in a small number of cases and that fires result from that, so to me it’s quite a plausible theory. If you want me to tell you exactly the fine detail of what was where and whether the flammable PVC wiring [insulation] inside the box was resting on the capacitor or whatever, I can’t because I don’t have that view but there are materials inside that enclosure that are flammable such as the PVC [insulation] on the cabling.[37]
[37]T.962.16-.27.
…
Counsel: So the capacitor fails and doesn’t breach the box, there’s a little fire in there which goes out relatively promptly; it would have to, wouldn’t it? --- I don’t know, the box may have melted, I don’t know. For reasons that I cannot tell you because I wasn’t there the fire didn’t progress at that time.
Counsel: No, but the air-conditioner continued to operate? --- Yes.
Counsel: If there was an explosion of the capacitor of the sort of energy you’re talking about there would be a loud explosion noise, wouldn’t there? --- There would be some noise, I agree with that, I can’t tell you that I have been in a place where a capacitor has exploded.
Counsel: But there would be a loud noise? --- There would be some noise, yes.
Counsel: So there would be a loud noise of some sort and you would certainly have the smell and the smoke? --- Yes, because the unit was operating so it would blow that smoke through the house.
…
Counsel: But there was no report of a loud noise? --- No, but she did point out that both Michelle and Mrs Jeffrey apparently woke at the same time at 4 a.m. in the morning and I presume they don’t sleep in the same room so one can speculate.
Counsel: The evidence was that Michelle got up and woke her mother? --- Was it, okay, I will accept that.[38]
[38]T.964.25-965.26.
…
Counsel: Is it your theory that after this explosion that something continued to be on fire or smoulder for six hours afterwards? --- I don’t know.
Counsel: You just don’t know? --- Yes.
Counsel: So you can’t really explain then all the smoke in the house? --- Well, the smoke in the house in my view came from the original event during the night which for some reason or another didn’t propagate to a full fire at the time but at 8 o’clock when the unit was turned on, the unit now being in a highly compromised state there was a fire that started at that time.[39]
…
Counsel: Your position now is, as I understand it, that this is the most likely theory? --- Yes, that’s correct.
Counsel: I suggest to you that this theory is completely far-fetched and the chances of it occurring are almost non-existent? --- I disagree.
Counsel: And the reason that you have moved back to this theory is because it’s occurred to you from reading Mr Kutek’s report that your other theories are equally as likely or unlikely? --- No, there is a rich history of capacitor
failing in air coolers and causing fires and since 2008 I have learned a lot more about air cooler and capacitor fires.[40]

[39]T.971.21-.31.

[40]T.972.18-.27.

Additionally, much of Dr Hart’s thesis ultimately found support in the concessions extracted in cross-examination of Mr Kutek. Despite the stridency with which he initially criticised Dr Hart’s analysis, ultimately Mr Kutek was led to confirm that there was a history of capacitors causing fires in air coolers:

Counsel: And can I ask you some questions, please, Mr Kutek, about the capacitors, you agree that they can cause fires? --- Yes.
Counsel: You agree they can start fires in air-conditioners? – If we mean air coolers, yes.
Counsel: You agree flammable gases from capacitors are ignitable? --- Yes.
Counsel: And you have seen and considered the various appendices to Dr Hart’s report which show a whole lot of products are recalled in relation to air-conditioners and other equipment that involve capacitor failures? --- Yes, I have seen those.
Counsel: You agree that provides a solid body of evidence of the Australian experience that these capacitors can fail and cause fires? --- It suggests that, yes.[41]

[41]T.1440.14-.28.

Mr Kutek also conceded that he had examined several fires where there was clear evidence of the failure of a component in an evaporative air cooler, including sometimes failure of a capacitor:

Counsel: In 2.2 in the fourth last line: ‘I have examined several fires where there is clear evidence that the failure of a component within an evaporative air cooler such as predominantly a pump motor or rarely a capacitor caused the fire’? --- Yes.
Counsel: So that’s something you have observed in fact in evaporative air coolers? --- Yes.[42]

[42]T.1446.13-.19.

Mr Kutek agreed, too, that when a capacitor ruptures it can expose the outer terminal and, albeit more rarely, the inner core:

Counsel: [Can] you assist the court and tell us what you have actually done, have you looked at lots of lots of these or a few? --- A few, less than 10.
Counsel: Thank you. They can rupture in a variety of ways? --- Yes.
Counsel: And if they are made of plastic their contents can rupture out the side of them? --- Yes.
Counsel: That would expose the metal terminal inside? --- It’s unlikely to expose the metal core but the outer terminal, yes.[43]

[43]T.1447.8-.15.

Mr Kutek accepted that, at relevant times, Seeley air coolers were manufactured with capacitors which had ferro-magnetic (scil steel) terminations:

Counsel: Can I take you over to 735E. Paragraph 2.11 you step into the breach and say: ‘In order to provide some firm evidence regarding the hypothesis that the steel strip was or was not from the main start capacitor I tested six that Seeley International indicated had been in use before during and after the time the manufacture of the Jeffreys unit, do you see that? --- Yes.

Counsel: That’s what you did, you did do the tests? --- Yes.
Counsel: So far as you documented them they are in either your commentary in this report or the photos attached? --- Yes.
Counsel: And you were able to confirm by reference to the is it Comar or Gomar (indistinct) capacitors, that two of them had magnetic termination? --- Yes.

Mr Kutek confirmed that the metal strip examined by Mr Moore and on which Dr Hart premised his thesis of the mechanism of ignition almost certainly came from an ICAR capacitor of the kind fitted to the Jeffreys’ Seeley air cooler:

Counsel (referring to a report prepared by Dr Hart): Next page, page 15, if I can direct your attention to the last paragraph, ‘because the ICAR MRL 25P type capacitor were used in earlier Breezair evaporative coolers, I am almost certain that the part Mr Moore inspected is from the motor capacitor of the Jeffery evaporative cooler? --- Yes.
…
HIS HONOUR: Do you agree with that proposition, do you think he’s right? --- If the premises are right then yes, he’s probably right.
…
Counsel: What I’m saying is here you cannot bring yourself to say Dr Hart is probably right that this metal component came from the capacitor inside this unit? --- I thought I just did say that.[44]

[44]T.1461.2-T1462.20.

Finally, when shown a report which he had prepared with respect to an earlier house fire caused by a Seeley Breezair air cooler in Lysterfield in 2004, Mr Kutek confirmed that that fire had been caused by failure of the capacitor showing very similar signs to those evident in this case:

Counsel: Then you go on and comment: ‘Such capacitors are usually housed in plastic cylinders and consist of metallized sheet insulation which is quite flammable’, and that’s true isn’t it? --- Yes.
Counsel: You say it is not unusual for such capacitors to be totally consumed during the fire? --- Yes.
Counsel: If that was the case one might find only the metal terminals and the metal contents of the plastic capacitor, mightn’t one? --- Correct.
Counsel: You see in the following paragraph 2.17 in the fourth line, reference to the experience of a very fine haze in the smoke in the hallway? --- Yes.
Counsel: And you prepared a second report in relation to the Lysterfield fire on 13 October 2004, if the witness could be given a copy of that please. If I can direct your attention please, to the first page, you see it’s dated 13 October 2004? --- Yes.
…
Counsel: There you ask the question: ‘What happened to the connecting wire and was it examined in detail before it disappeared?’, do you see that? --- Yes.
Counsel: Then you make this proposition: ‘Any evidence of electrical damage to the connecting wire or to the terminations of the capacitor would have provided reliable evidence that the capacitor was the probable cause of the fire’? --- Yes.[45]

[45]T,1456,11-T. 1458.15.

Counsel for Seeley complained that the judge was in error in accepting the dancing arc theory given that Dr Hart had not tested it and was reliant on a text book as support for it and because, it was said, Dr Hart had not seen the capacitor housing and did not know whether or not the required explosion would have ripped the housing apart.

We do not think that either of those criticisms is warranted. Starting with the dancing arc, and bearing in mind that Mr Kutek accepted that electrical arc damage to the capacitor metal strip would have provided reliable evidence that the capacitor was the probable cause of the fire,[46] it is to be observed that the reason which Mr Baghurst first gave in evidence in chief for rejecting the possibility of the dancing arc theory proposed by Dr Hart was that he could not conceive of a capacitor with some kind of void sufficient to facilitate arcing by electric conduction through a gaseous medium:

Counsel: And you have read [Dr Hart’s] evidence about the dancing arc theory? --- Yes I have, yes.
Counsel: Is the dancing arc theory, what are your comments on that theory as postulated by Dr Hart, do you regard that as a theory which is possible? ---
…
HIS HONOUR: I assume you strongly disagree with that theory? --- I do, yes.
Counsel: Can you say why? --- Yes.
Counsel: As [a] matter of theory can you say why? --- Yes, I can, because in order to – even to start thinking about the possibility that this happened, one needs to postulate that the capacitor is actually – some kind of container with a void within it. The capacitor is not a void and as I said just a few moments ago, it’s in fact a thing which is solid, there is nothing – there is no atmosphere, there is no gas, no spaces within these capacitors….it’s a solid cylindrical thing which is mainly dielectric, that is mainly plastic material, with a small amount of distributed but not uniformly distributed aluminium, then a whole element is I would say potted [or] encapsulated inside a tube which is sometimes plastic, sometimes metal, and then the whole lot is sealed off at the top with again a hard epoxy or similar resinous material so there are not gases, spaces whatever within such a capacitor. The notion of a dancing arc, an arc of course by definition being the electrical conduction through a gas. There’s no gas within this capacitor, there is no gas within the capacitor unless the capacitor has been blown apart.[47]

[46]T.1458.11-15 and T.1462.3-.20.

[47]T.1518.12-1519.16.

At that point, the judge reminded Mr Baghurst that the hypothesis was that the capacitor had blown apart, so that there could well have been a void. Mr Baghurst then gave as his reason for rejecting Dr Hart’s thesis that there could be no arc without two exposed electrodes in close proximity to each other and that, if the capacitor exploded, it would not blow away or destroy the shrouding of the inner electrode so as to expose it:

HIS HONOUR: Well, I think that is what Dr Hart is hypothesizing … we have one of these blow out incidents of some description and it is that which removes the state you have just mentioned where everything is enclosed and therefore there can be no question of any dancing arc. We have metal exposed after the blowout. I think that’s the starting point for the theory… --- Yes, and I certainly accept that. There are a number of other problems, however, the first is that – an essential ingredient to arcing is to in this particular case, two metallic electrodes.
…
Counsel: But in this case? --- … So the notion that you could take a capacitor even if, as your Honour suggests, that capacitor has been blown apart, … Any blowing apart of the capacitor … is not going to blow away or destroy as I see it, the shrouding of that inner electrode and therefore I cannot with the best and most independent thought processes that I can bring to bear on this subject, conceive of a dancing arc within the capacitor.[48]

[48]T.1519.26-T.1522.13.

When later cross-examined, however, Mr Baghurst ultimately conceded that, if the capacitor had exploded, as it could have done, and melted the PVC insulation on wires within the control box in which the capacitor was mounted, which was certainly possible, there could have been arcing between the central strip of the condenser and the exposed wiring:

Counsel: Indeed if you could look at exhibit 8 for me, which is a photo of the capacitor in the housing that was produced by Seeley International in the trial? --- Yes.
Counsel: You see there that capacitor which is clipped into place held there? --- Yes.
…
Counsel: If the wiring was exposed to a fire the plastic insulation of the wiring could melt, couldn’t it? – Yes, it’s probably PVC and PVC is flammable, yes.
Counsel: If you had an explosion of some kind and this was a plastic PO [type condenser of the kind in fact fitted] as opposed to an aluminium capacitor, the remnants would be held in position, wouldn’t they, by those clips? – Yes, and also the shape of the cover which conforms.
Counsel: Exactly, and so in that environment rather than having your novel theories of something you get in a toy shop where you have the dancing arc, you have an immediate proximity between these wires and the remnants of this PO capacitor, don’t you, that’s the environment in which this accident could happen? --- It certainly is, yes.
Counsel: So if one looks at and demystifies the dancing arc however [un]fortunate the use of that expression was, it’s not difficult for you as an electrical engineer to understand the very possibility of those wires once the PVC’s been melted of by a fire being exposed to the remnants of the metal terminal inside that capacitor, is it? --- I think this is a very, very long --- You think it’s a long bow? --- yes, I do.
…
Counsel: Notwithstanding your initial report about the significance of this metal strip and your apparent certainty at that stage it had a lot to do with the fire? --- I have never said that I don’t regard this [as] a scenario in which any of those phenomena could possibly take place.[49]
[49]T.1629.30-T.1631.7.

(iv) Exploding capacitor

Turning next to whether it was likely that the capacitor might have exploded in the fashion suggested by Dr Hart, it seems to us that there was really no dispute about it. So far as we can see, the evidence from all available sources implies that it was distinctly possible that the capacitor exploded in just the fashion which Dr Hart envisaged.

In evidence in chief, Mr Baghurst readily accepted that capacitors do blow apart and cause fires:

Counsel: Do they blow themselves apart, Dr Hart talked about an explosion which he said was not necessarily a Hiroshima event, is that what occurred, you have tested these two [sic] to fail, haven’t you? --- Yes.
Counsel: Many times? --- Yes, I have.
Counsel: How do they fail? --- Well, in all kinds of ways, sometimes the case will simply split apart and you end up with well, the capacitor as presumably you have seen around here with a split in it and that’s it and it stops operating. Sometimes there’s a little bit of transient flame. I have never had to put out a capacitor by dousing it with any fire extinguishing medium or anything like that. Sometimes the end which carries the terminals will be blown clear and the body of the capacitor then is in one place and I would call it the encapsulate of the ceiling, a material which is more often than not some kind of resinous material still carrying the terminals is somewhere distant from the capacitor, maybe half a metre away or so, still carrying the connecting leads and there might be a transient flash and maybe - - -
HIS HONOUR: This is a failure open, and once the failure has occurred the flow of current stops so there is no danger of any fire? --- I think it would be a very brave person who said there is no danger of any fire.
HIS HONOUR: Dr Hart said on a number of occasions in effect everybody knows capacitors can cause fires, and I’m just wanting to understand is what you have just been telling me a refutation of that or do you accept it or what’s your view on that point? --- I think I would have to say practically all equipment that handles energy of any kind is subject to fire and I couldn’t possibly say that – I couldn’t possibly refute the fact that it’s probable to cause a fire….[50]

[50]T.1516.21-1517.18.

In cross-examination, Mr Baghurst also accepted that the reason that capacitors fail in the field is because a series of breakdowns may occur closely spaced within the capacitor causing the temperature to rise and destroy the protective film. He agreed, too, that, if that occurred in a P0 capacitor (which was the kind of capacitor fitted to the subject air cooler), there was at least a chance of failing in a dangerous manner:

Counsel: Yes, and look immediately to the right-hand side of that commentary from the ERA publication about the failure of capacitors is: ‘The plastic film at the breakdown point is carbonized with a release of gas. A series of breakdowns occur close together within the capacitor or closely spaced in time may not be able to dissipate the heat generated by the breakdown events. Temperature then rises locally with each breakdown, gas accumulates, sufficient breakdowns, temperature rises high enough to melt the plastic film’? --- Yes.
Counsel: You agree generally, don’t you, with the substance of what this article states, don’t you? --- I certainly do, yes.
So notwithstanding that a design feature of this plastic or metallized film wrapped thousands of times around the core are designed to self-heal, the fact is they are known to fail by the mechanism described in this article? --- That is mechanism by which they fail in the destruction test too, yes.
Counsel: And in the field, look at figure 2? --- Yes, all electrical equipment failed in the field.
Counsel: As you I [sic] said yesterday the most expensive car in the world will hit a brick wall? --- Yes.
Counsel: There we are. But the PO is going to be the car that’s designed with no safety failure? --- I think it has the least [protection against] chance of failing in a dangerous manner.[51]

[51]T.1614.28-T.1615.21.

Mr Baghurst accepted as well that, after the subject air cooler was manufactured, an Australian Standard came into force (although it has since changed) which required air coolers of that type to be fitted with a P2 type capacitor.[52] The significance of that was that the P0 type was a base grade capacitor which had no safety protection in the event of failure. Above that, in terms of quality, sat the P1 type which could fail in either open circuit or short-circuit but was protected against fire or shock hazard. Above that again stood the P2 type, which was the best available and the only one of the three which was both designed to fail only in open circuit and protected against fire or shock hazard.[53]
[52]T1503. 20-.1504,2 and T.1612.12-T.1613.12 and T.1617.6-T.1618.10.
[53]T.1611.15-1614.12 and 1617.6-.31.

Additionally, there was the evidence of Mr Kutek, earlier referred to, that there was a solid body of evidence of Australian experience that capacitors within evaporative air coolers of the kind in question can and do fail causing flammable gases to escape and ignite and cause fires, and also the evidence of the Seeley senior technical officer in the previous proceeding earlier referred to that there can be arcing between foil layers within the capacitor which progressively consumes the capacitor by driving off the fire retardants and allows the polymer to oxidise resulting in the rapid propagation of fire.

(v) Testing a ‘very similar motor’

Mr Baghurst claimed in his evidence in chief that he had tested ‘a … very similar motor’ to the motor fitted to the subject air cooler and thereby demonstrated that the motor in the subject air cooler could not have started with the condenser failed in short circuit.[54] He said that in his opinion that showed that Dr Hart’s thesis was fundamentally flawed. In counsel’s submission, the judge erred in failing to explain why he rejected that evidence.
[54]T.1523.11-1559.29.

In our view, that submission faces difficulties at several levels. To begin with, as was demonstrated in Mr Baghurst’s cross-examination, the test which Mr Baghurst purported to have carried out on a ‘very similar motor’ was in fact carried out on a ‘generically similar’ motor which he had been told was ‘very similar’ to the subject motor.[55] There was no independent evidence as to how similar the motors were, still less why the test was not carried out on a motor which was identical to the subject motor.
[55]T.1635.24-T.1636.4.

Next, the control box fitted to the test motor was different to the control box fitted to the subject motor and there was no evidence that the capacitor within the control box was the same P0 type capacitor which was fitted to the subject control box.[56] Mr Baghurst said he was not concerned with the control box. But, since it contained the capacitor, it is difficult to think why that should be so and he did not condescend to explain why it should be so.
[56]T.1635.13-.23.

Next, the only test which Mr Baghurst said he conducted was to bypass the capacitor in its entirety. He did not test it, as he agreed he could have done with a variable condenser in the space of about 10 minutes, in a fashion which would much more closely have approximated to a partial short circuit of the kind postulated by Dr Hart.[57] The test was not video-recorded and, although it was said to have been witnessed by a member of Mr Baghurst’s staff, that member of staff was not called to give evidence.
[57]1638.11-T.1639.1.

Additionally, although Mr Baghurst claimed at the outset that the test demonstrated it was impossible for the subject motor to have started if the capacitor had failed in closed circuit, later he backed down to the point that the highest he could put it was that it was ‘highly unlikely’.[58] Then, in response to questioning by the judge, he disclosed for the first time that when he tested the motor by short circuiting the capacitor, the motor in fact turned at about 60 revolutions per minute.[59] That does not seem to us to be inconsistent with Dr Hart’s thesis that he expected that ‘some degradation of motor performance would have occurred however, this may not have been noticeable to the occupants’.
[58]T.1639.7
[59]T.1570.22-.31.

There was also good reason to doubt Mr Baghurst’s impartiality, and it is apparent that the judge did doubt it. As was earlier remarked, in his first report Mr Baghurst readily accepted that the metal strip may have been arc damaged (as Mr Moore had determined by metallurgical microscopic examination to be the case) and, therefore, that it was likely to have been causative of the fire. At that stage, Mr Baghurst was contending that the metal strip could not have come from the Seeley air cooler because, according to him, Seeley air coolers contained no ferro-magnetic parts. Contrary to that assertion, it subsequently emerged that Mr Baghurst well knew as at 2004 (although he said he had forgotten about it by the time of his first report) that Seeley used P0 capacitors in their air coolers,[60] and that he had carried out destruction tests for Seeley on P2 capacitors which Seeley was then proposing to introduce in place of ICAR P0 capacitors of the kind fitted to the subject air cooler, which had a central metal strip.[61]
[60]T.1621.9-.12.
[61]T.1623.5-.23.

Consequently, in his second report, Mr Baghurst was obliged to accept that the metal strip found at the scene of the fire had come from the capacitor in the Seeley air cooler. In order, however, to overcome the conclusion which would have followed that the air cooler was thus the cause of the fire, Mr Baghurst then advanced for the first time a new thesis that the metal strip could not have caused the fire because the metalized plastic layers surrounding it would have been incapable of sustaining the dancing arc. Of course, that contention faced the difficulty that Mr Moore had already determined by microscopic examination that the metal strip was arc damaged. Hence, seemingly in order to get over that problem, Mr Baghurst withdrew his concurrence with Mr Moore’s assessment that the metal strip was arc damaged and alleged for the first time that he had now determined by viewing some photographs of the metal strip[62] that the damage to it was not arc damage but rather the effects of aluminium deposited from the capacitor and combining with the steel in the steel strip in order to form a ferro-aluminium alloy.
[62]The strip itself having by then been lost.

The judge took up the significance of those changes in Mr Baghurst’s opinion by way of several questions which his Honour asked of Mr Baghurst as follows:

HIS HONOUR: So we started off with a situation where Mr Moore says I have found the smoking gun, here it is, it’s a metal strip, and the first response is … who can take any notice of Mr Moore, he thought it was copper and it turns out it’s really iron, so he has no credibility. The next response well, actually, and in any event this wasn’t part of the unit so even if this is the smoking gun it’s the smoking gun implicating something other than the Breezair, then, well, it is actually part of the unit, well maybe it’s not really a smoking gun at all; isn’t that really the way matters have progressed? --- Yes.

HIS HONOUR: And you can see how I think [counsel for the respondent] in due course is going to say that the defendant’s experts approached this matter not on the basis of giving the best evidence and in the advancement of justice but acting as advocates for the acquittal of the air-conditioner…
HIS HONOUR: Assuming that, what do you say to that? --- I say that as it stands that’s a pretty fair sort of summing up of the situation. I think what – I would really take exception with [is] the notion that the expert witnesses, for want of a better word, are necessarily partisan, I take - - -
HIS HONOUR: I didn’t expect that you were going to agree with that but I thought in due course [counsel] was going to invite me to make that finding so I thought I would give you the opportunity to respond to it? - - - Thank you, Your Honour …[63]
…
HIS HONOUR: Am I right in saying the reason you regarded Mr Moore’s theory as lacking in credibility when you wrote that [first] report was because he traced everything back to a metal strip that you believed you had conclusively established was unconnected to the air conditioner and that was the point his evidence lacked credibility …? --- Yes, and of course later I was astounded it should later be indentified as steel, not copper.
HIS HONOUR: Yes, which we mention again so I again, not wanting to steal [counsel’s] thunder but he’s going to say well, once that deficiency was made good you have got to go back, the only reasonable thing for you to do now is say okay, Mr Moore, now we know that strip did come from the cooler I have to agree with you so why don’t you have to do that? --- I agree with him it formed part of the cooler, what I now understand … [is that it is] abundantly clear that the erosion on that strip is almost – was almost certainly caused by alloying with aluminium which also forms part of the capacitor and which is a notion he rejected…
…
HIS HONOUR: Which sounds a lot more dramatic. [counsel] has just put to you’re not a metallurgist so the identification of this bite out of the strip as caused by alloying is a key step in your conclusion? --- Yes.
HIS HONOUR: As we now see it but are you qualified to take that step based on your expertise? --- I believe I am, yes, I don’t call myself a metallurgist I underwent a five year course in engineering and that included a year subject called engineering materials.[64]

[63]T.1602.18-T.1603.16.

[64]T.1605.27-T.1607.6.

Ultimately, the judge did not refer to that exchange in his reasons for judgment. But his Honour’s observation as to Mr Baghurst not explaining why he found it ‘quite unrealistic’ to suppose that the motor could have started after the capacitor had failed in closed circuit appears to us to imply that the judge did not think much of Mr Baghurst as an expert witness. The judge also remarked on the ‘caustic’ criticism which Mr Baghurst had heaped on Mr Moore and Dr Hart’s opinions and, although his Honour did not refer as such to the fact that most of those criticisms were ultimately shown to be misplaced, there is no reason to doubt that his Honour also took that into account.

(vi) Dr Hart’s change of opinion

Ironically, counsel for Seeley submitted that it was Dr Hart’s expert testimony which was suspect – because over time Dr Hart advanced three different theories of causation – and counsel argued that the judge erred in failing to refer to the impact on Dr Hart’s credibility of those changes.

We reject the argument. If anything, it seems to us that, in view of the circumstances which caused Dr Hart to make successive changes to his opinion, the changes added weight to his conclusions.

It will be recalled that, in his first report of 20 August 2008, Dr Hart advanced five possible scenarios of causation, including as Scenario 5 that the motor start capacitor failed and caught fire. At that stage, he discounted Scenario 5 because he did not know, and could not have known, that the metal strip found at the scene of the fire (which Mr Moore had determined to be arc damaged) came from the core of the start capacitor.

By the time of his second report, which was dated 7 April 2010, Dr Hart had been afforded the opportunity of inspecting an electric induction motor of the type used in the subject air cooler and an example of the ICAR P0 type start capacitor included in the subject cooler capacitor enclosure. Even at that stage, however, it remained problematic as to whether the metal strip assayed by Mr Moore had come from the subject cooler capacitor. In those circumstances, Dr Hart concluded:

… I cannot exclude fire causation by capacitor failure. The action of turning on the cooler at 8am may have been the final straw for a capacitor that had internal damage.

Then came Dr Hart’s third report, which was dated 19 April 2012, prepared after and because of the late revelation by Mr Baghurst that he had discovered that the external and central electrodes of the ICAR Ecofill capacitor model MLR225 PRL were ferro-magnetic and after receipt of information provided by Mr James of Seeley that a very similar ICAR Ecofill model MLR 25P capacitor was used in the subject air cooler. As Dr Hart explained in his third report, that information prompted him to dissect an ICAR Ecofill capacitor which he had obtained during investigation of house fires started by Brivis evaporative air coolers, and as a result he found that the shape of the central metal terminal strip was similar to the strip shown in Mr Moore’s photos, albeit shorter because the MLR225 PRL was slightly shorter than the MLR 250 model.[65] Thus, Dr Hart was able to conclude with near certainty that the metal strip which Mr Moore had inspected and determined to be arc damaged was from the motor capacitor of the subject air–cooler:

Because the ICAR MLR 25P type capacitor were used in earlier Breezair evaporative coolers, I am almost certain that the part Mr Moore inspected is from the motor capacitor of the Jeffrey evaporative cooler.[66]

[65]AB 3, C421 [160], [170].

[66]AB 3, C426 [240].

Coupled with further information provided by Mr James of Seeley, which revealed with equally near certainty that the metal strip assayed by Mr Moore was the central electrical connection from inside an ICAR motor capacitor, that implied that the electric arc damage observed by Mr Moore was highly significant:

… Mr James revealed that the electrical enclosure is made from moulded GE Noryl 90 plastic. I interpret this statement to apply to the main control box …
Noryl 90 has a softening temperature of about 110°C. In the event of a capacitor fire the plastic enclosure whether it be Noryl 90 or not, will soften and flow. The fire could escape from the plastic capsule.
…
The information from Mr James raises the possibility that the 5W ceramic resistor was on the circuit board inside the capacitor enclosure. This is consistent with the strip that Mr Moore examined being from the capacitor because both parts were found in the debris of the air cooler.
I agree with Mr Kutek when he says that ‘the steel strip may have the same origin as the resistor and circuit board’. That is, from inside the capacitor housing.[67]
...
I have concluded that the metal terminal strip that Mr Moore inspected is almost certainly the central electrical connection from inside an (ICAR) motor capacitor and that this capacitor was used in the Jeffrey air cooler.
I believe that the damage to the strip that Mr Moore examined is highly significant. Mr Moore observed that there were two areas of damage on the metal strip. … I believe that the damage to the strip that is evident in Figure 4 is due to clashing between two electrically live metal parts; that is short-circuit. The result is localized melting/erosion of metal due to intense heating…[68]

[67]AB 3, C428, [260]-[280].

[68]C430, [310].

In turn, as Dr Hart explained, all those things when combined led to the conclusion that the most likely cause of the Jeffrey fire was that the motor capacitor failed and caught fire:

Opinions About Fire Causation
In previous reports I considered several fire scenarios and excluded some…
I believe that the strip is the central electrode of the motor capacitor. I believe that the damage to the strip is electrical short-circuit damage and that this indicates that the capacitor was live when some other electrode or wire touched the central strip. … there must have been an explosive failure of the part for this damage to occur. The short-circuit probably involved direct contact between the central strip and the other terminal metal.
…
In my opinion the most likely cause of the Jeffrey fire is that the motor capacitor failed and caught fire. The fire could not be contained within the enclosure on the motor, either because the plastic material softened and flowed resulting in a hole, or because it was punctured by the explosive [force]. The causal factor of the failure was probably ag[e]ing of the polypropylene insulating film. The contributing factors were probably capacitor quality, continuous operation without cooling airflow and maybe relatively high electricity-supply voltage.[69]

[69]AB 3 C 431-434.

Seen in that context, there is nothing about Dr Hart’s changes of opinion or the fact of the judge acting on them which detracts from their efficacy. Nor do we accept that the judge was bound to say anything more about them than his Honour did in his reasons. It is obvious that Seeley did not provide timeously all of the information which it should have and, when that information was later revealed, Dr Hart had to take it into account.

(vii) Dr Hart’s specific changes of mind

Finally, counsel for Seeley submitted that Dr Hart’s evidence involved three specific important and unexplained changes of mind which the judge failed to take into account. The first was said to be that, whereas in his initial report Dr Hart stated that he could not conclusively determine whether the damage observed by Mr Moore was arc damage, in his final report he stated that he was convinced that it was.

The judge did not fail to take that into account. The reasons for the change were fully explained in Dr Hart’s final report, some of which is extracted above. As Dr Hart there set out in detail, upon examination of the MLR25 PRL capacitor and after receiving the further information from Mr James of Seeley, it became apparent that the metal strip was the central electrode of the capacitor and that it was damaged in two areas in two different ways. That concomitance of circumstances rendered it highly likely that one of the two areas, the one shown in Figure 4, was arc damaged.

The second change of mind was said to be an unexplained alteration in Dr Hart’s estimate of the volume of electrical current required to rend the capacitor. In his first report, Dr Hart stated that to cause arc damage of the kind shown in Mr Moore’s photographs would have required a current of hundreds of amps and that such a current would likely have tripped the circuit breaker before the damage in question could occur. Later, Dr Hart changed his mind about that and concluded that the damage could have been inflicted by a ‘dancing’ or ‘moving contact’ arc along the strip which would have gradually eroded the strip without tripping the circuit breakers.[70]
[70]T.938.8-T.939.21.

Contrary, to counsel’s submission, that change of mind was not unexplained either. Dr Hart explained it thus:

… This capacitor is connected to the start winding circuit so if you have a connection here that is making and breaking and moving, whatever, you will find that you will get sparking continuously and it will gradually erode the strip away, so I distinguish that mechanism from having a 100 amp type fault that only lasts for a fraction of a second.[71]

We also note that counsel for Seeley did not put to Dr Hart any reason why that analysis should not be accepted.

[71]T.939.15-.21.

The third change of mind was one from the conclusion which Dr Hart stated in his first report that Scenario 5 was unlikely (because, if there had been a failure of the capacitor during the night, the motor could not have started again at 8.00 am in the morning) to the conclusion expressed in Dr Hart’s final report that it would have been possible for the motor to start again at 8.00 am if the capacitor had failed in closed circuit.

As has already been observed, the explanation for that was that, at the time of writing his first report, Dr Hart had been led to believe that the subject capacitor was of P2 type and, therefore, could only fail in open circuit. If it failed in open circuit, it could not have conducted the current required to start the motor again at 8.00 am. In contrast, by the time of the final report, which was later confirmed by viva voca evidence, it had been established that the subject capacitor was an ICAR type P0 capacitor which could have failed in closed circuit. Hence, as Dr Hart explained in his evidence, the closed circuit was capable of providing the kind of current required to restart the motor albeit that he would have expected the motor to perform poorly and to have given off an unusual noise.[72]
[72]T. 965.1-.30.

Finally, counsel for Seeley criticised what he said was the inconsistency between the statement in Dr Hart’s initial report that there was no report of abnormal operation of the air cooler when it was started up at 8.00 am and his evidence in the witness box that he did not know what the Jeffreys said about the operation of the cooler at that stage.

We do not think those two propositions to be inconsistent. Believing that there was no report of malfunctioning and not knowing what the Jeffreys said about it are not inconsistent. And in any event, consistently with Dr Hart’s thesis that the motor would likely have been affected by the capacitor and not functioning as normal, Mrs Jeffrey’s evidence was that when she switched the cooler back on at 8.00 am she ‘noticed that it[73] was worse’ and ‘I noticed that it had a bit of an electrical burning smell’.[74]
[73]A reference to the smoke haze.
[74]T.299.12-.15.

Ground 2 – Downlights

Under the heading of Ground 2, counsel for Seeley criticised the judge for rejecting as unlikely that the downlights in the ceiling caused the fire. In counsel’s submission, his Honour was wrong to place emphasis on the fact that there was no explanation of why (if the fire were the result of the downlights causing the roof insulation to smoulder and burn) such a fire did not occur before 2003; particularly given that Mr Kutek gave evidence of having seen such an occurrence (of a fire starting long after such roof insulation was installed in a home) and Mr Baghurst said that he could conceive of the possibility.

In our view, the submission is misplaced. The fact that there may have been other instances of fires starting years after installation of roof insulation did not gainsay that, in this case, there was no evidence as to why, if the fire were the result of the lights causing the insulation to smoulder, it did not happen before 2003. Since there was no suggestion, still less credible evidence of what it was about the insulation which might have altered in the 15 years since it was installed, the fact that it had been installed for 15 years without incident was logically a significant, relevant consideration in excluding the lights as the source of the fire.

It was also submitted that the judge erred in attributing any or as much significance as his Honour did to the fact that photographic evidence of the ceiling in the vicinity of four of the five downlights showed no signs of the kind of discolouration which was said to be typical of downlight induced conflagration. In counsel’s submission, the judge failed to deal with Mr Kutek’s and Mr Ranzenhofer’s evidence that such discolouration might not always occur or with the fact that there was no photographic evidence of the ceiling proximate to the fifth downlight.

In fact, however, the judge dealt at length with the significance of there being no evidence of the condition of the ceiling near the fifth light and concluded that there was nothing in the evidence or otherwise which contradicted
[75]Reasons, [126].
Messrs Ranzenhofer’s and R’s opinions, based on their examinations of the ceiling space, that the ceiling lights were not a possible cause.[75]

Counsel also argued that the judge gave undue emphasis to the fact that Mr Jeffrey was able to step onto the metal roof of the house as he attempted to put out the fire issuing from the air cooler. In counsel’s submission, that was a vague piece of evidence because there was nothing to say where exactly Mr Jeffrey had walked on the roof and no evidence as to where the air cooler ducts were located.

We think that criticism to be captious. All that his Honour said about the matter was to make the observation, en passant, that:

[Mr Jeffrey] was able to move about the roof during the early stages of the fire in his bare feet. This tells against any fire having ‘flared up’ outside the footprint of the aircooler.

There is no error in that. It was not put to Mr Jeffrey that he did not go where the heat would not have travelled from the air cooler. Nor was he asked anything in that connection concerning the placement of the air cooler ducts. In the circumstances, the judge was quite entitled to treat the matter as he did. We would do the same.
Ground 3 – Circumstantial case reasoning

Lastly, under Ground 3, it was contended that the judge failed to follow the correct process of reasoning for a circumstantial case of finding the facts based on the evidence, determining whether on the whole of the circumstantial evidence it could be concluded on the balance of probabilities that it was the start capacitor which caused the fire and then considering the expert evidence based on the facts as found. According to counsel for Seeley, the only finding of fact which was open on the evidence was that the air cooler was working normally on the morning of the fire. Thus, it was said, the judge erred in finding, contrary to the evidence, that the Jeffreys were wrong when they said that the air cooler had worked normally.

We do not accept that submission either. Among other problems, it faces the difficulty that the Jeffreys did not say that the air cooler worked ‘normally’ when they turned it back on at 8.00 am. Mr Jeffrey said only that he heard it ‘click’ or ‘kick’ on and that he did not recall whether any air was being blown by the fan into the bedroom.[76] Mrs Jeffrey said that when she turned it on ‘it was working’ but then ‘I noticed that it had a bit of an electrical burning smell not like a wood fired smell and so I told Noel and I turned it off’.[77] Thus, as the judge found:

… It is quite conceivable that the sound of the ‘click’ or ‘kick’ of the motor may not have been the sound of the proper and normal functioning of the motor, but rather of something abnormal and sub-optimal, perhaps with more resistance involved than Professor Baghurst’s experiments would have entailed…. On that hypothesis, the sound made was abnormal but, through lack of attention, the Jeffreys failed to notice.[78]

[76]T.213.14-.19.

[77]T.299.8-.14.

[78]Reasons, [154].

Perhaps, by itself, Mr and Mrs Jeffrey’s evidence would not have warranted a finding one way or the other as to whether the air cooler worked normally when it was turned back on at 8.00 am. But that was not the end of the matter. As has been seen, the judge was required to stand back and view the picture from a distance, making an informed, considered, qualitative appreciation of the whole of the evidence, including Mr and Mrs Jeffrey’s evidence of what they recalled of the cooler’s functioning in the circumstances which obtained at the time. Fundamental among the objective aspects of the whole of that evidence was that, within minutes of the air cooler being turned back on, it emitted a burning electrical smell, causing Mrs Jeffrey to turn it off, and then almost immediately started to billow smoke before bursting into flames. Viewed as a common sense question of fact,[79] it was probable that the air cooler was the culprit.
[79]March v StramarePty Ltd (1991) 171 CLR 506, 515 (Mason CJ), 523 (Deane J).

Ground 4 – Findings against the weight of the evidence

Under Ground 4, counsel for Seeley contended the judge erred by adopting a process of false reasoning to the effect that, having rejected other suggested competing hypotheses as unlikely, it was open to conclude that the capacitor was the cause of the fire. So to reason, counsel argued, was to engage in a process of choosing between mere possibilities which was tantamount to speculation. In counsel’s submission, even on the view of the evidence most favourable to the Jeffreys, it was not open to the judge to reach a real level of satisfaction that the capacitor was to blame.

We do not think that the judge erred in his process of reasoning. It is clear enough from his Honour’s judgment that he was cognizant of the need to be satisfied on the balance of probabilities and to avoid conjecture and surmise. As his Honour said:

I was referred to a number of recent authorities on the issue of standard of proof in civil proceedings. It is fair to say that they raised some arcane questions that arose out of some fairly arcane fact situations. In my view the best guidance as to standard of proof here is an old case determined by a distinguished court Bradshaw v McEwans Pty Ltd.[80] The court was Dixon, Williams, Webb, Fullagar and Kitto JJ and in a joint judgment their Honours said inter alia:
... this is a civil and not a criminal case. We are concerned with probabilities, not with possibilities. The difference between the criminal standard of proof in its application to circumstantial evidence and the civil is that in the former the facts must be such as to exclude reasonable hypotheses consistent with innocence while the latter you need only circumstances raising a more probable inference in favour of what is alleged. In questions of this sort where direct proof is not available it is enough in the circumstances appearing in the evidence give rise to a reasonable and definite inference: they must do more than give rise to conflicting inferences of equal degrees of probability so that the choice between them is mere matter of conjecture .... But if circumstances are proved in which it is reasonable to find a balance of probabilities in favour of the conclusion sought then though the conclusion may fall short of certainty it is not to be regarded as a mere conjecture or surmise ... All that is necessary is that according to the course of common experience the more probable inference from the circumstances that sufficiently appear by evidence or admission, left unexplained, should be that the injury arose from the defendant’s negligence. By more probable is meant no more than that upon a balance of probabilities such an inference might reasonably be considered to have some greater degree of likelihood.[81]

For the reasons given in the findings which I have indicated above the hypothesis as to the role of the motor capacitor satisfies the criteria laid down by their Honours in the quoted passage. Bradshaw’s was a case where the evidence available at trial was in their Honours’ words ‘meagre’. Due to the extraordinary circumstances surrounding this case which I have explained already, whilst the trial was of an inordinate length and there was an overwhelming volume of evidence, as to the most crucial matters in light of the widespread destruction caused by the fire and the failure of proper documentation in the immediate aftermath and before the removal of the remains, on the crucial questions of causation the evidence was nevertheless ultimately as meagre as was the evidence in Bradshaw’s case.[82]

[80](1951) 217 ALR 1.

[81](1951) 217 ALR 1, 5, 6.

[82]Reasons, [158]-[159].

Nor do we accept that the evidence was insufficient to establish on the balance of probabilities that the capacitor was the cause of the fires. Possibly, his Honour’s reference to the evidence on the ‘crucial questions of causation’ being ‘meagre’ could be seen in isolation as suggesting an inadequacy in proof of the ultimate issue. But the observation needs to be understood in context. As we read it in context, it is confined to the precise mechanism of capacitor failure as opposed to the fact of capacitor failure. And, as it appears to us and seems to have presented to his Honour, there was substantial circumstantial evidence of the fact of capacitor failure.

Apart from anything else, it was clear beyond peradventure that, when the air cooler was switched back on at 8.00 am on the day of the fire, it resulted within minutes in the smoke in the house becoming thicker, the emission of an electrical smell from the unit and then the power to fail. Mrs Jeffrey sensed that there was ‘something terribly wrong’ and almost immediately she could see that smoke was billowing from the air cooler. Moments later, it burst into flames. Subject to other competing hypotheses, those facts implied that the air cooler had failed.

Contrary moreover to the submissions advanced on behalf of Seeley, so to reason was not to resort to post hoc propter hoc syllogism but rather to bring to bear on a common sense issue of fact the reality of everyday life experience. Logic and common sense imply that, if one turns on an electrical appliance and, within moments, it results in the emission of smoke, the production of an electrical smell, the mains power to fail and then smoke and flames to issue from the appliance, there is a fair chance that the appliance has suffered an electrical breakdown.

Thirdly, arc damage to the central metal strip of the capacitor as determined by Mr Moore’s metallurgical examination of it made it highly likely that the central strip of the ICAR P0 capacitor fitted to the Jeffreys’ air cooler sustained electric arc damage. Overall, Mr Kutek’s evidence was supportive of that conclusion. Mr Baghurst’s claim to the contrary was rightly rejected. He had no expertise in metallurgy and he had only ever seen a digital photograph as opposed to the strip itself. Nor was it disputed that the metal strip could not have suffered arc damage unless the capacitor had failed.

Fourthly, the likelihood of the capacitor having failed in one way or the other was enhanced by Mr James’ and Mr Kutek’s evidence of previous failures of similar capacitors in service and Mr Baghurst’s evidence that capacitors can and do fail in all sorts of ways.

Fifthly, as has been explained, the judge rightly rejected each of the suggested competing hypotheses as unlikely possibilities.

Sixthly, it may be that the judge was unable to conclude on the balance of probabilities that the mechanism of failure was a ‘dancing arc’. As we have said, we think that to be the significance of his Honour’s reference as to the evidence of the mechanism of failure being ‘meagre’. But it was unnecessary for the judge to be persuaded of the exact mechanism of failure. It was enough to be persuaded that the capacitor had failed in closed circuit and, to that end, for it to be shown that there was at least one possible means by which that could have happened; particularly given that, according to Mr Baghurst’s evidence, there were potentially many other modes of capacitor failure with P0 capacitors which would not have occurred if the capacitor had been of the P2 variety.

Finally, counsel for Seeley argued that it was not enough for the judge to reject the suggested alternative hypotheses of downlight ignition and toroid, transformer and triad. In counsel’s submission, it was conceivable that there were other unidentified possible explanations for the fire and, because the judge had not identified them, let alone concluded that the evidence or the absence of it was sufficient to reject them, his Honour’s acceptance of the capacitor as the cause of the fire remained an exercise in conjecture.

We think that submission to be misplaced. So far from it being an exercise in conjecture to eschew consideration of supposed but unidentified competing hypotheses, it would have been quite improper to take them into account. As has been explained, even in a criminal case it is unnecessary to do more than exclude reasonable hypotheses consistent with innocence; and in a civil case such as this, a tribunal of fact is entitled to be persuaded of what is alleged by identifying the circumstances which render it more probable than not.[83] More precisely, a trial is not an inquiry into the truth of every fact and circumstance in the sequence of events under consideration. It is a rational adjudication of the issues inter se. And it is to be determined according to the pleadings. Where, therefore, it is alleged that the cause of a catastrophe was one thing and the defence is that it was another, the outcome should, in all but exceptional cases, be decided on the evidence according to which of them is the more likely.
[83]Nom v Director of Public Prosecutions [2012] VSCA 198, [112]; and see The Hon Mr Justice D H Hodgson, The Scales of Justice: Probability and Proof in Legal Fact-finding, (1995) 69 ALJ 731, 747.

Conclusion

In the result, the appeal will be dismissed.

- - -

[2013] VSCA 288

18 October 2013