Veolia Water Central Ltd v London Fire & Emergency Planning Authority

This is a dispute about the financial responsibility for the maintenance of fire hydrants in the London area. The Claimant, who was formerly known as Three Valleys Water plc (and to whom I will refer as "Three Valleys"), is a statutory water undertaker whose area of responsibility includes North and North West London. The Defendant, to whom I will refer as "London Fire", is the Fire and Rescue Authority for London.

The principal issue in the case is what is meant by the obligation to keep fire hydrants in "good working order". Three Valleys, as a water authority, has a duty under section 57(3) of the Water Industry Act 1991 ("the Act") to keep fire hydrants on its mains in good working order. By section 57(3) of the Act its expenses of doing so are recoverable from the local Fire and Rescue Authority. The position taken by Three Valleys is that any leak from a fire hydrant, however minor, should be repaired. Otherwise, it says, it will not be complying with its duty under section 57(3) of the Act to keep fire hydrants on its mains in good working order. London Fire takes a different view. Its case is that fire hydrants will be in good working order if they can be used for fire fighting purposes without any material restrictions or dangers. It submits that minor leaks, and similar defects, do not affect the firefighters’ ability to use fire hydrants for fighting fires and consequently that they should be under no obligation to pay for them to be rectified.

As a result of this difference of approach London Fire has refused to pay for the costs of carrying out repairs to rectify what it considers to be minor leaks, or similar trivial defects, that have been invoiced by Three Valleys between 21 June 2002 and 17 September 2007. Three Valleys claims that it has incurred expenses in carrying out repairs in order to keep fire hydrants in good working order to the extent of £527,299 (including VAT but excluding interest), being the work to which those invoices relate. London Fire contends that all these repairs were carried out without its authority and that they were unnecessary to keep the hydrants in good working order.

Three Valleys issued a claim form on 21 January 2008 claiming the sum that I have mentioned, and this is the trial of four preliminary issues that were ordered to be tried by Master Rose on 17 February 2009.

So far as this case is concerned, there are two principal types of hydrant, both of which are installed in a hydrant pit with a metal frame and cover. The first is known as a Type 1. This hydrant has a valve in the nature of a sluice gate. The gate itself is wedge shaped in section and the end is circular (as viewed along the axis of the pipe). Pressed into the rim of the gate is an annular face ring made of gunmetal which forms the seal with the faces of the ends of the pipes. The gate is hollow and accommodates a threaded spindle which, when turned, raises and lowers the gate. At the top of the casing which houses the gate and the threaded part of the spindle shaft there is a collar which prevents the spindle from moving upwards. The top of the housing is known as the bonnet. As the spindle is turned the gate is pushed down until it forms a tight seal against the upstream and downstream faces of the pipe. This also forces the collar up against the underside of the bonnet.

The spindle passes through the bonnet and terminates at the top in the shape of a truncated obelisk of square cross section. Onto this is fitted a key which turns the spindle. Where the spindle shaft passes through the bonnet there is an annular space which is or used to be packed with a lubricated hemp packing (this has now been replaced with packing made of some form of PTFE). The purpose of the packing is to act as a seal to prevent any water that enters the cavity under the bonnet from escaping. When this happens it is known as a spindle leak.

There was an issue between the experts as to the extent to which, if at all, the cavity below the bonnet would become filled with water, possibly at the same pressure as the mains. In theory this should not happen when the valve is closed because when the spindle is screwed fully home the gate should form a watertight seal all the way round the face of the pipes. However, Three Valleys’ expert, Mr Whittaker, said that if the gate is not manufactured so as to form a perfect fit, or perhaps as a result of years of wear, there would be some very slight play. The effect of this would be that when the valve is exposed to mains water at full pressure, the gate will be pressed firmly against the face of the downstream pipe possibly leaving a small gap between the upper parts of the seal and the face of the upstream pipe. This small gap would enable water from the pipe to leak into the cavity below the bonnet. I will have to return to this issue later.

Of course, when the gate is raised the seal will be broken in any event and water will flow into the cavity under the bonnet. However, once the valve is fully open the water can be discharged through the hydrant outlet and into the fireman's hose.

Type 1 sluice valves can leak in one of two ways when closed (or apparently closed). There can be a leak past the upper part of the gate seal into the cavity under the bonnet, which can then give rise to a leak past the spindle if the packing that I have described is not forming a good seal, or there can be a leak via the gate seals into the outlet, or perhaps both. The evidence in this case suggests that in some circumstances the leak through the spindle can be very violent as the hydrant is opened. This is known as a “gusher”. Sometimes it may take the form of a fairly full flow of water from the top of the hydrant, at other times it may take the form of a strong jet of water issuing vertically from the spindle housing. It is common ground that the latter is an obvious source of danger to those using the hydrant, particularly if small pieces of grit and the like are carried up with the water.

A leak through the gate into the outlet will gradually fill the outlet unless the water can drain away through the frost valve, which is fitted to the underside of the outlet chamber or pipework near its lowest point. However, the frost valve will only permit water to drain away at a fairly low rate and so, if the leak is anything more then very minor, water is likely to fill the outlet pipework. If the leak is even more serious then the water will overtop the outlet, fill the hydrant pit and then overflow. The pit can also fill up with rainwater at times of heavy rain and this may well mean that possibly contaminated surface water can enter the outlet of the hydrant. One of the problems presented by a leaking gate valve is that if there is a sudden depressurisation of the mains, water from the outlet can be sucked back into the main. If the water sitting in the outlet is dirty or contaminated this will discolour or contaminate the mains water (which should, of course, be wholesome and fit for drinking).

Although it is theoretically possible that if a hydrant pit was to fill up with rain water to a level above the bonnet, water could also be sucked back through the spindle shaft in the event of a sudden depressurisation of the main, this seems to me to be a very remote possibility because the high water level in the pit would probably be fairly transitory and any depressurisation of the main would have to occur whilst this high water level in the pit remained. I am not aware that anyone has suggested this as a relevant risk.

The other type of hydrant is a Type 2 hydrant. This is usually installed on a vertical branch of pipe so that its inlet connection is usually at its base. This hydrant has a screw down plug valve which prevents the flow of water by being pushed down into a seating area just above the inlet connection. This type of hydrant is also operated by a threaded spindle. The Type 2 hydrant has a similar gland packing arrangement to the Type 1 hydrant, and its spindle has a similarly shaped top.

The plug valve of the Type 2 hydrant is made of metal and coated with elastomer and in the older models, with which I am concerned in this action, it has a circular disc or washer of rubber attached by a nut and a securing washer (in order to facilitate replacement of the washer). When the washer is in good condition the valve will achieve a good seal. If there is not a good seal, then the hydrant presents similar problems to those discussed in paragraphs 9 and 10 above in relation to the Type 1 hydrant.

Both types of hydrant require what is known as a false spindle cap. This is a hollow cap of the same profile as the top of the spindle and it fits over it in order to protect the relatively soft bronze of which the spindle is made. The key to operate the hydrant is fitted on the spindle cap. It appears that spindle caps are relatively attractive items of scrap metal and that from time to time they are stolen.

The sample claims with which I am concerned have been chosen to reflect the five main categories of repair that have been carried out on Type 1 and Type 2 hydrants. They are:

There is no dispute about the fact that over time the gland packing around the spindle can deteriorate and provide a path for water to leak out along the shaft of the spindle. Re-packing is fairly common and is carried out when a spindle leak is detected. A major area of dispute is the severity of the leak that has to be present before repacking is reasonably justified or, more accurately, gives rise to a right of recovery of the costs of doing so from London Fire. Over the period of the sample claims repacking was usually carried out using some sort of PTFE material, instead of oiledhemp. Sometimes this material would suffer from what is known as stress relaxation, which takes place when after being subjected to pressure it does not return to precisely its former shape. This means that it may no longer achieve an efficient seal and so the housing bolts may have to be tightened in order to rectify this.

Re-washering is carried out to Type 2 hydrants to rectify a leak through the main valve. There is again a dispute about the severity of leak which must exist before re-washering is reasonably justified.

Replacing a defective hydrant occurs in the circumstances described above.

Fitting a false spindle cap is a simple operation if that is all that is required. However, it appears that sometimes a damaged spindle cap becomes jammed onto the top of the spindle so that a fairly elaborate operation is required in order to remove it. This may involve demolishing part of the sides of the pit in order to gain the necessary access. However, this type of repair is not the subject of any of the sample claims.

The replacement of a frame or cover will be necessary if the existing frame or cover has come damaged or distorted and either presents a hazard to the public or prevents the hydrant cover from being readily removed. It is obvious also that firefighters must be able to lift the cover without delay when they need to use a hydrant.

The preliminary issues that have been ordered to be tried are the following:

Although there are only 11 sample jobs, there are some 1460 outstanding invoices issued by Three Valleys which London Fire has refused to pay. They cover the period from 18 June 2002 to 30 August 2007, but the dispute is ongoing. As I have already said, the sum claimed in respect of these invoices is £527,299 (including VAT but excluding interest).

Since section 57 of the Act is central to the issues in this case, I must set it out in full. It is as follows:

“57 Duty to provide a supply of water etc for fire-fighting

(1)

It shall be the duty of a water undertaker to allow any person to take water for extinguishing fires from any of its water mains or other pipes on which a fire-hydrant is fixed.

(2)

Every water undertaker shall, at the request of the fire and rescue authorityconcerned, fix fire-hydrants on its water mains (other than its trunk mains) at such places as may be most convenient for affording a supply of water for extinguishing any fire which may break out within the area of the undertaker.

(3)

It shall be the duty of every water undertaker to keep every fire-hydrant fixed on any of its water mains or other pipes in good working order and, for that purpose, to replace any such hydrant when necessary.

(4)

It shall be the duty of a water undertaker to ensure that a fire and rescue authority has been supplied by the undertaker with all such keys as the authority may require for the fire-hydrants fixed on the water mains or other pipes of the undertaker.

(4A) Where a fire-hydrant is removed (other than at the request of the fire authority concerned) by a water undertaker in the course of carrying out works in relation to any of its water mains or other pipes, the cost of replacing the fire-hydrant shall be borne by the undertaker.

(5)

Subject to subsection (5A) and section 58(3) below, the expenses incurred by a water undertaker in complying with its obligations under subsection (2) to (4) above shall be borne by the fire and rescue authority concerned.

(5A) Where a fire-hydrant is damaged as a result of any use made of it with the authority of a water undertaker, other than use for the purposes of fire-fighting or for any other purposes of a fire and rescue authority, the fire and rescue authority is not liable for the cost of repairing or replacing the hydrant.

(6)

Nothing in this section shall require a water undertaker to do anything, which it is unable to do by reason of the carrying out of any necessary works.

(7)

The obligation of a water undertaker under this section shall be enforceable under section 18 above by the Secretary of State.

(8)

In addition, where a water undertaker is in breach of its obligations under this section, the undertaker shall be guilty of an offence and liable –

(a)

on summary conviction, to a fine not exceeding the statutory maximum;

(b)

on conviction on indictment, to a fine.

(9)

In any proceedings against any water undertaker for an offence under subsection (8) above it shall be a defence for that undertaker to show that it took all reasonable steps and exercise d all due diligence to avoid the commission of the offence.”

There is also a separate and different duty imposed on water undertakers by section 37 of the Act. This provides as follows:

“37 General duty to maintain water supply system etc

(2)

It shall be the duty of every water undertaker to develop and maintain an efficient and economical system of water supply within its area and to ensure that all such arrangements have been made –

(a)

for providing supplies of water to premises in that area and for making such supplies available to persons who demand them; and

(b)

for maintaining, improving and extending the water undertaker’s water mains and other pipes,

as are necessary for securing that the undertaker is and continues to be able to meet its obligations under this Part.

(3)

The duty of a water undertaker under this section shall be enforceable under section 18 above –

(a)

by the Secretary of State; or

(b)

with the consent or in accordance with a general authorisation given by the Secretary of State, by the Director.

(4)

The obligations imposed on a water undertaker by the following Chapters of this Part, and the remedies available in respect of contravention of those obligations, shall be in addition to any duty imposed or remedy available by virtue of any provision of this section or section 38 below and shall not be in any way qualified by any such provision.”

For many years fire authorities have implemented a system of inspecting fire hydrants in their regions at regular intervals. In the case of London Fire, this used to involve a visual inspection of each hydrant every year and a physical test of each hydrant every other year. If these inspections or tests revealed any defect, then (from about 2001 onwards) London Fire’s own hydrant technicians would repair the defect if they could. But this would only be possible in the case of relatively minor defects that did not require isolation of the mains. If a more serious defect was found, London Fire would issue a request to the water undertaker to carry out the necessary repairs. The water undertaker would then repair the defect and charge London Fire the cost of doing so. London Fire would inspect the hydrant in question and, if satisfied that the repairs had been properly carried out, would pay the charge.

As the Act makes clear, the water authority was only entitled to recover from the fire authority its actual expenses incurred in carrying out the repairs, and was not permitted to recover anything in respect of profit. This is common ground. What is not common ground is the extent to which, if at all, the charges that Three Valleys seeks to impose include an element of profit (or expense not associated with cost of the repairs) rather than just direct overheads.

In about 2001 London Fire changed its system of inspections by discontinuing the annual visual inspection and instead confining the regime to the carrying out of a test on each fire hydrant every other year. At this time water undertakers did not carry out any systematic inspections of fire hydrants in their areas. However, in about late 2002, Three Valleys and other water undertakers began to respond to pressure, in particular, from the regulator Ofwat, to reduce the extent of leakage from water mains. They instituted programmes of systematic checking of water mains in their areas for leaks, or at least stepped up existing such programmes. If leaks were found, they were repaired.

It seems that fire hydrants and similar outlets were a common source of leaks and so, when Three Valleys discovered a leaking fire hydrant, they would request London Fire (or any other fire authority within their region) to issue an order for the repair of the leak. However, London Fire was not prepared to issue orders for the repair of leaks or other minor defects in the absence of any evidence that the leak in question was something more than a minor leak (or some other defect) that would or might impair its firefighters’ ability to use the fire hydrants for fighting fires.

So London Fire's position, then and now, is that the statutory obligation under section 57(3) of the Act does not extend to remedying defects that do not affect the fire fighting capability of a hydrant. However, it is prepared to pay (but purely as a concession, it says, rather than as an obligation) for repairs to hydrants or to the chamber in which they are located that are necessary to prevent the hydrant installation being a danger to the public. For example, if the iron cover to a hydrant pit is broken such that it may present a tripping hazard to members of the public, then London Fire is and would be prepared to pay for the cost of its repair.

Consistently with this position, therefore, London Fire has refused to pay for minor repairs to hydrants, mainly in respect of minor leaks, for which it has not issued an order. In fact it has gone further, because it has taken the view that it should not have to pay for any repair that it did not order unless it has evidence that the repair was necessary to render the hydrant fully operational. Prior to 2002 this problem seldom arose because Three Valleys did not carry out any inspections of fire hydrants on a regular basis, if at all. Where Three Valleys became aware of a defect at a hydrant, perhaps because a member of the public or local authority had reported it, they would notify London Fire in the first instance unless the need for repairs was urgent. But once it became necessary for Three Valleys and other water undertakers to carry out systematic checks to detect leaks, its leakage detection teams would check hydrants attached to any section of the mains in which a leak was suspected. From this time onwards the approach of Three Valleys was that any leak, however minor, should be repaired. It was this approach that brought it into conflict with London Fire and which eventually gave rise to this claim.

London Fire accepts that Three Valleys may well be under a duty to repair minor leaks, but under section 37 of the Act, not under section 57. The repair of minor leaks may ensure that a system of water supply is economical (as section 37 requires), but it is not necessary - says London Fire - to repair such leaks in order to ensure that fire hydrants are kept in good working order.

This dispute has been the subject of discussions and correspondence between the parties since 2003, but they have been unable to reach any measure of agreement. London Fire's position as it was advanced in this case was succinctly set out in a letter dated 30 June 2004 (E1/798) from John Anthony, the Assistant Commissioner of London Fire, to Mr Steve Ross, the Head of Customer Operations at Three Valleys. Mr Anthony wrote as follows:

“Firstly, let me state that the role of all Fire Authority’s [sic] is to reduce loss of life and damage to property from fire. In London our objective is stated very simply - to make London a safer city. In terms of securing water for fire fighting, we undertake a biennial inspection programme of all 108,000 fire hydrants across the capital. Where any hydrant is found to be inadequate for its specific purpose (that is, for the purpose of permitting a supply of water to be taken to extinguish fire) and where we are unable to perfect minor repairs in house, we are happy to place orders for repair or exchange as necessary. Over and above this, we are happy to place orders for hydrant repairs where the hydrant is either unable to operate in accordance with its specific purpose or poses a significant hazard to either the user or the public. We believe that these policies allow us to meet our statutory and stated obligations to the people of London and permit the efficient and responsible allocation of public resources into this area of our work.

The advice we have received from Counsel is quite clear in that Fire Authority’s [sic] are not responsible for all hydrant repairs. Section 57 -2 [sic] of the Water Industry Act states that “a hydrant is for the purpose of permitting a supply of water to be taken to extinguish any fire that has broken out in the vicinity". Sections 57(3) and (5) of the same Act further makes clear that Fire Authority’s [sic] are not liable for the costs of all fire hydrant repairs, but only those that require them to be fit for the purpose, again ". . . to permit the hydrant to be used in order to extinguish any fire in the vicinity". Drawing water from a fire hydrant is only undertaken in a limited number of all fire calls that we respond to and we are satisfied that our rolling inspection programme and the orders placed arising from that fully meet our obligations in maintaining an efficient and economical fire and rescue service.

. . .

. . . In our experience, the significant majority of leakage from fire hydrants is minor, does not affect the use of the fire hydrant for its stated purpose and in many cases is a result of the hydrant not having been properly closed following use. Our own fire crews are well versed in the operation of fire hydrants and the importance of closing them after use, as you would expect of a professional emergency service whose responsibilities include not only fire but attending flooding incidents also, manipulation of the hydrant through opening is often sufficient to remedy minor leaks. In any event, we are satisfied that responsibility for such leakage does not fall to the Fire Authority to fund under this section of the Act".

I would make one comment straight away about this letter. The second of the paragraphs that I have quoted above does not in my view summarise correctly the effect of the provisions of the Act. Section 57(2) requires the water undertaker to fix fire hydrants, at the Fire authority's request, "at such places as may be most convenient for affording a supply of water for extinguishing any fire which may break out within the area of the undertaker". It therefore seems to me that the reference to "affording a supply of water for extinguishing any fire" is primarily directed to the location of the fire hydrants rather than to the purpose of each fire hydrant itself.

Later in 2004, following fairly protracted discussions, Three Valleys and London Fire concluded a Memorandum of Understanding (“MoU”) in relation to the policy to be adopted towards the repair of fire hydrants. The MoU was signed on the half of London Fire on 1 September 2003, and on behalf of Three Valleys on 19 August 2004. It will be apparent from what I have already said that by this time the dispute in relation to the cost of repairing minor leaks was already on foot.

The MoU recited that its purpose was to "to form a common approach and a basis for establishing an efficient working relationship between the two parties". It was expressly provided that the document was not a contract and was not legally binding.

However, since both parties have referred to the MoU, particularly Appendix 4, I should set out its provisions. They are as follows:

“4.1

The Fire Authority shall be responsible for the cost of maintenance of all accepted installations, except where damage is caused by a hydrant user licensed by the Water Company or other circumstance as agreed.

4.2

All hydrant defects found by the Water Company and/or reported to them by members of the public will be notified to the Fire Authority as soon as is reasonably practicable. The Fire Authority will be advised where any defect renders the hydrant inoperable.

4.3

Repairs to the fire hydrant, which are likely to cause an interruption to supply or require a shut-off, shall only be carried out by the Water Company or its approved contractors.

4.4

When inspecting or testing a fire hydrant the Fire Authority may, at their own cost, repair minor defects, (such as damaged or missing tablets and plates, minor pit defects or other minor repairs that have no impact on the quality of water in the main) and shall report defects (not repaired), giving rise to danger or leakage, to the Water Company as soon as is reasonably practicable. Defect reports will be confirmed by the issue of an official order for the repair.

4.5

Where it is considered necessary to carry out work without an order or additional work over and above that detailed on the official order, the Water Company will, where practicable, make reasonable endeavours to advise the Five Authority in advance. All such work will be notified in writing to the Fire Authority as soon as practicable.

4.6

Upon receipt of an order from the Fire Authority, the Water Company will complete the necessary work endeavouring to meet the following times.

i dangerous/priority work requiring immediate attention - 24 hours

ii urgent work requiring quick attention - 14 days

iii routine work- 60 days

Examples of works that may fit the above categories are as follows:

i any defect that represents a significant safety hazard to the Fire Service or the public.

ii any defect that has rendered the hydrant unusable for fire fighting purposes

iii any minor defect that could result in the hydrant becoming unusable

4.7

. . .

4.8

Charges for repairs or replacement fire hydrants shall be in accordance with Appendix 10.

4.9

The Fire Authority shall be informed in writing that the fire hydrant is available for inspection within 14 days of repair or replacement.

4.10

As soon as practicable and within 28 days of receipt of the written notice of repair, the Fire Authority shall inspect the fire hydrant and sanction payment upon receipt of a valid invoice or inform the Water Company that the repair is unsatisfactory, stating its reasons."

On 2 September 2004, about two weeks after Three Valleys had signed the MoU, Mr Anthony wrote again to Three Valleys - this time to Mr Jeff Bishop - in the following terms (E1/838):

“I undertook to write to you further following our meeting on the 20th August. I agreed that I would set out the reasons for the Authority's position in the light of the advice received from Counsel.

Firstly, the question of minor leakage from fire hydrants that does not affect their use for their stated purpose (that is, to provide a supply of water for the purpose of extinguishing fire) has considerable financial implications for this Authority given the 108,000 fire hydrants installed across the capital. More significantly, to allocate taxpayers’ money to an area of work that does not in any way contribute to providing a more efficient fire and rescue service would be an inefficient and unacceptable allocation of resources.

. . .

. . . The fire authority is liable to meet the cost of repair and maintenance where (a) the hydrant is not working at all, (b) where a hydrant is working but not to a standard that is fit for its stated purpose and (c) where reasonable routine maintenance is being carried out for the purpose of ensuring that a hydrant is fit for the purpose of providing the water required to extinguish any fire. We do not believe that any minor leakage that we discussed at our meeting falls within these categories and that in the majority of cases such leaks have no effect on the operation of the fire hydrant for its stated purpose.”

One thing that was clear from the evidence of Mr Lewis, who was Three Valleys’ network service manager for the North London East region, was that when one of Three Valleys Customer Service Teams visits a hydrant in response to a report of a leak they do not make a judgement, and are not required to make a judgement, as to whether or not the leak is going to impair the functionality of the hydrant fighting fires in the next two years or so (see Day 3, at page 2). Rather, Mr Lewis accepted that the approach taken by Three Valleys was that where a leak was found it had to be repaired (Day 3, at 45).

There was much cross examination of the witnesses on each side about this and other contemporaneous documents and, whilst this was helpful in clarifying why the parties had adopted the positions that they have now taken, it has been of limited assistance in relation to the questions that I have to decide. However, it did assist me to form views about the reliability and objectivity of the evidence given by the witnesses. At this point I should record my impressions of the four principal witnesses, and the two experts, who gave evidence before me.

Mr Lewis was Three Valleys’ network service manager for the North London and Eastern region. He was responsible for repair and maintenance operations in the region, and he had control of approximately 50 staff, being a mixture of direct Labour and sub contract labour, with five network managers. He has spent some 35 years in the water industry.

He was called to explain the differences between the two types of fire hydrant with which the case is concerned, and he gave evidence in relation to the sample claims with which I have been asked to deal. Much of his evidence on this aspect was little more than an analysis of Three Valleys’ records since, perhaps inevitably, he had little if any personal knowledge of the work carried out to the various hydrants concerned. However, during cross-examination in particular, he gave a certain amount of evidence which was evidence of opinion and therefore outside the scope of his role as a witness of fact. Whilst I do not doubt that he was well qualified by experience to give expert evidence in certain areas, there are good reasons why expert evidence is not usually admitted from witnesses of fact – the principal ones being that they may not be independent of the parties to the dispute and that they are not exposed to the rigour and discipline and the very strict rules that apply to the evidence of experts.

But leaving aside the fact that he was not called as an expert, and that in consequence his opinion evidence was technically inadmissible, I formed the view that much of that evidence was of limited value. This was either because much of it appeared to be little more than a restatement of TVW’s position in the litigation, or because it was couched in fairly general terms: for example, when referring to possible damage from a continuing leak, he said "however if the defect is left over a prolonged period then damage can be done to the bottom seating of the hydrant and in such cases could lead to the hydrant needing to be renewed" (witness statement, paragraph 17.3, at B/154). In saying this, I do not mean to suggest that he was seeking to be unhelpful, only that he had probably become too immersed in this litigation to be able to give an objective opinion on some of the matters in issue.

However, in relation to one or two of the sample claims, Mr Lewis was prepared to make concessions in cross examination against Three Valleys’ interests and that was entirely proper.

Mr Ross is a civil engineer who between 1997 and 2002 was responsible for the North London region of Three Valleys network, later taking the Eastern region as well, when he was responsible for all aspects of work relating to the company's network and principal customer management in that area. In late 2002 he took over the role of head of customer operations. I found him to be a careful witness, whose evidence I felt was generally reliable and consistent with the contemporaneous documents. Where he did not know the answers to questions, or where matters were not within his own knowledge, he was prepared to say so. Although on one or two occasions it was suggested that he had changed his evidence - and there were indeed some differences in both substance and impression - I did not find those differences to be of any real significance.

However, in relation to the details of Three Valleys system of charging for overheads his evidence turned out to be of limited value. His role was limited to using figures that others had prepared and given to him and into which he had no input.

He had worked for the London Fire Brigade for about 25 years, in a number of different administrative roles. For the past 9 years he has been the Water Manager for London Fire. He has never been an operational firefighter, although he had attended Mr Lewis's Master Class on fire hydrants. He said that although he had witnessed the recruit training course, he had no training as a firefighter. Whilst he gave his evidence in a very precise manner I was not confident that it was always reliable. For example, at paragraphs 10-12 of his witness statement he described London Fire’s Policy and Procedure Document for Hydrants and he explained in some detail the system of testing hydrants and the various outcomes that would be reported on the inspection sheet. There was nothing in this part of this statement to suggest that this was no longer current policy; indeed the document was put to Mr Lewis in cross examination (Day 2, at page 69) with no indication that the system of using the four levels of defect referred to in paragraph 12.2 (E1/553) had been abandoned fairly shortly after the document was produced. However in the course of his evidence Mr Hanson was referred to certain inspection sheets which did not appear to have been completed following the system set out in the Policy and Procedure Document, to which his response was that he thought that this was one of the reasons why London Fire removed these levels. He said that it was "because they were posing some confusion" (Day 5, 69). The following day he told the court that this change in practice had been in place for some time (Day 6, page 72).

I was left with the impression that Mr Hanson was another witness who has been rather too close to this litigation to be able to maintain an objective approach and, as a result, I have reservations about the reliability of some aspects of his evidence.

He is the Assistant Commissioner for Operational Policy for London Fire. He has been with the London Fire Brigade for 30 years, starting as a firefighter and working his way up through the operational side. Although his evidence was fairly limited in scope, I found him to be an impressive witness.

Mr Ian Whittaker, the expert instructed on behalf of Three Valleys, is a materials engineer employed by Yorkshire Water Services Ltd. In that capacity he is the engineer responsible for standards and product specifications for the procurement of products used in water distribution. He is or has been a member of a large number of British Standards committees concerned with various aspects of the water industry.

He was understandably cross examined by Mr Michael Curtis QC, who appeared for London Fire with Ms Deok Joo Rhee, to the effect that he was not a truly independent witness because his employer, Yorkshire Water, could be adversely affected financially if the decision in this case were to go one particular way. Mr Whittaker accepted this. In addition, Mr Whittaker accepted that he has for some years effectively acted as a spokesman for the water industry in various capacities.

I accept that for these reasons Mr Whittaker is not a completely independent witness. However, I consider that in spite of this he was very conscious of his duties as an expert and did his best to prevent his inherent conflict of position from affecting his professional judgement. Nevertheless, he came to the case inevitably with views that reflected his background which I consider that no amount of intellectual detachment could effectively put aside. To this extent therefore, I consider that his evidence needs to be scrutinised with some care, although I must make it quite clear that I intend no disrespect or criticism of Mr Whittaker in saying this.

Dr Sanders is a Principal Member, which I take to be the equivalent of a partner, in Burgoynes, the well-known consulting scientists. He is a chemical engineer by background. For about the past 12 years has specialised in the forensic investigation of incidents, mostly involving fires or escapes of water from various systems. It was unsatisfactory that his CV was not attached to his report as it should have been (and which I have still not seen).

Dr Sanders candidly admitted that he had no practical or theoretical experience of fire hydrants and, indeed, he thought that he may not have set eyes on one until he was instructed in this case. In fact, it is probably not unfair to say that at the time when he was instructed he may well have known less about fire hydrants than any other witness. However, I thought that he gave his evidence fairly and he was prepared to accept that some of his initial views required some changes.

In the event, the two experts produced a very helpful Joint Note and I did not find that there were any very significant differences between their views. Somewhat unsurprisingly, they tended to differ on matters of emphasis rather than substance. However I should deal with the principal issues on which they differed.

I have already mentioned that there was an issue between the experts as to the extent to which the cavity under the bonnet in a Type 1 hydrant would be filled with water under pressure (it is common ground that the packing glands do not come under pressure in a Type 2 hydrant). For this to happen there must be a leak between the upper part of the upstream face of the gate and the corresponding face of the pipe. Mr Whittaker took the view that in older Type 1 hydrants there would be water under pressure in the cavity under the bonnet so that the gland packing round the spindle would also be under pressure. Dr Sanders disagreed: his view was that the cavity under the bonnet, and hence the gland packing, would only become pressurised if the valve gate was slightly open.

There was a short video showing the opening of a Type 1 hydrant at 4 Kingsmead Drive. What the video showed was an open pit containing a Type 1 hydrant, largely filled with mud or silt but leaving the spindle cap accessible, which was then slowly opened. As the key was turned about 1/8^th of a turn water started to flow into the pit from the area of the spindle. This continued for about 10-15 seconds and then the spindle was turned a further quarter turn, at which point the rate of flow of water increased dramatically (such that it started to flow over the edge of the pit). This continued for about five seconds and then suddenly stopped, at which point the standpipe (which was fitted with a pressure gauge at its outlet) came under full pressure, which was about 3 bar.

The experts interpreted this sequence of events in different ways. Mr Whittaker said that the initial turn of the key was not sufficient to activate movement of the gate but was sufficient to release the force between the spindle thrust collar and the bonnet so that water could flow through to the spindle gland. Since the gland packing was not achieving a good seal, water was able to flow through it as observed in the video. As the key was turned further, the gate was unseated allowing more water into the cavity under the bonnet, which then increased the rate of the observed flow of water into the pit, until the gate and spindle assembly were lifted by the water pressure thereby partially re-engaging the thrust collar and the bonnet.

The view of Dr Sanders was that the initial turning of the spindle started to release the gate, allowing flow past the gate into the cavity under the bonnet (and presumably into the outlet also) giving rise to the observed flow of water. Further opening released the gate fully, initially increasing the leak rate and then stopping or substantially slowing it as the spindle was pushed upwards and the collar engaged under the bonnet.

These two accounts show that one thing about which the experts apparently agree is that the increase in water pressure in the cavity under the bonnet can cause the spindle assembly to be pushed upwards thus causing the collar to engage with the bonnet and thereby stop or reduce the flow of water through the gland packing. The difference between them is the effect on the valve gate of the first turn of the key. It appears to be common ground that the spindle of the Type 1 hydrant has what is known as "slack angle": that is the degree of play in the thread of the spindle that allows the spindle to be turned very slightly without actually lifting the gate. Having watched the video I doubt very much that the initial 1/8^th turn of the spindle could have both disengaged the gate and absorbed the amount of slack angle, whereas I can see that the further quarter turn would have released the gate from its seat. The respective views of the experts indicate that when the collar is fully engaged under the bonnet it forms a seal, if not a perfect one. This suggests that if any mechanical pressure on the collar was relieved, that seal would be broken (provided that the upthrust from the pressure of the water in the cavity under the bonnet was not sufficient to keep the collar in place).

In my view, Mr Whittaker's explanation of the events in the video is to be preferred. It seems to me unlikely that such a small initial rotation of the spindle could have released the gate from its seat. This leads me to conclude also that Mr Whittaker is probably correct in his view that the cavity under the bonnet is likely to be pressurised in the Type 1 hydrants. So on this point I accept the evidence given by Mr Whittaker, both in his supplemental report (C/224, paragraph 4.3) and when he was cross examined on 28 October 2009 (Day 7, at pages 45-58), that the cavity under the bonnet of a Type 1 hydrant is usually pressurised and as to the reasons for it. It follows from this that I reject the suggestion made by Dr Sanders that the presence of a leak from the spindle implies that there will also be a leak through the gate.

Whilst London Fire accepts that a "gusher" requires repair, it submits that minor spindle leaks and other minor leaks and drips do not need to be repaired in order to keep a hydrant in good working order. There has been no attempt to define precisely what constitutes a minor leak or drip, and that is something that I will have to attempt to define - assuming, of course, that it becomes necessary to do so.

The more difficult question, raised by Three Valleys, is whether it is always necessary to repair a minor leak because there will always be a real risk that it will get significantly worse within the two year inspection interval.

This has become a potentially important issue because, in a very elegant and well crafted closing submission, Mr Javan Herberg, who appeared for Three Valleys, sought to demonstrate on the basis of what was stated in the experts Joint Note that there was a significant or real risk that a minor spindle leak could develop into a gusher, or at least to an extent that would result in water filling and overflowing the pit, within the two year interval between hydrant inspections. On this basis, he submitted, the repair of a minor spindle leak would always be justified.

In order to do justice to this argument I need to set out the contents of sections 1.5 and 1.10 of the Joint Note in full. They are as follows:

Three Valleys’ case on this aspect (which was in the alternative to its primary case that no leakage can ever be permitted) was that, even on London Fire’s approach, a hydrant with any form of spindle leak requires work to keep it in good working order, within s.57(3) of the Act. This was for one of two broad reasons: where there is a spindle leak, there is a real or significant risk that within the relevant time period (being the 2 year inspection interval) a minor leak could develop into a gusher, or at least to a leak such that would result in water filling and overflowing the pit.

An important step in the argument was the submission that by “worsen significantly”, the experts were jointly identifying the time which it would take for a minor or potential leak to be of the nature of a “gusher”. This, submitted Mr Herberg, was confirmed by section 1.10 of the Joint Note, because there the experts’ view about the likelihood of becoming gushers was expressly linked back to the discussion of rate of deterioration at section 1.5 of the Joint Note. He submitted also that this was the basis on which Dr Sanders gave evidence in cross-examination (on Day 7, pages 148–151).

I do not accept that by agreeing that a minor spindle leak would deteriorate slowly, taking months or years to "worsen significantly", the experts were agreeing that these more serious leaks would have reached a degree of severity that equated to a gusher or a leak that would overtop the pit. That may or may not have been the view of Mr Whittaker, but what is recorded in the Joint Note does not in my view go that far – particularly when read with the comments of Dr Sanders in relation to the sample cases (whether those reasons were right or wrong is irrelevant for this purpose). Further, I do not consider that Dr Sanders made such a concession under cross examination. The relevant passage in the transcript is as follows:

At the time, I did not understand the "yes" given by Dr Sanders in answer to Mr Herberg's first question in this extract to be agreement with the first sentence of the question, but merely an acknowledgement in response to the last part of the question, namely that that was the sort of thing they were talking about. That impression is in my view consistent with his answer as recorded in the transcript.

In re-examination, Dr Sanders said that the significant worsening contemplated might lead to the pit being overtopped or that it filled up more, but that this did not necessarily mean that this would give the Fire Brigade a problem in using the hydrant, as he understood London Fire's evidence (Day 7, pages 190-191). Thus, he said, he did not intend the term "worsen significantly" in the Joint Note to mean that a leak would worsen to a point where a hydrant was compromised for fire fighting.

In my view this evidence, taken by itself, falls well short of establishing that there was a real or significant risk that within two years a minor spindle leak could develop into a gusher. As to whether a minor spindle leak might lead to the overflowing of a pit, the position is less clear. I do not regard this evidence as establishing that there was a real or significant risk that within two years a minor spindle leak could become a leak of severity sufficient to cause the overflowing of a pit. I took Dr Sanders to be conceding that a pit being overtopped was one of the things that might happen within “months to years”, but not that there was a significant risk of it happening before the hydrant was next inspected.

There was evidence, which I accept, that minor leaks could sometimes be stopped by the simple expedient of closing the hydrant properly. Mr Hanson gave an example (at B/119) of one such leak at a hydrant at 15 Wolsey Close, Southall. Three Valleys had notified London Fire on 14 August 2006 that this hydrant was leaking. The hydrant was inspected by Mr Hanson who found that the leak stopped as soon as the valve was closed properly. The request for a repair order was then withdrawn. Mr Lewis also said that manipulating the hydrant's spindle could stop a spindle leak, although the next time the hydrant was operated the leak would probably re-appear (B/146). However, when giving evidence in chief Mr Lewis said that, although the leak had been effectively stopped, it would get worse because the valves were so old (Day 2, at pages 25-26). When challenged about this in cross examination (Day 3, at pages 89-90), Mr Lewis accepted that the leak would only start again when somebody next interfered with the hydrant, unless there were substantial differences between the night time and daytime pressures (which he accepted was no more than a theory). He then said that "The apparatus is designed not to leak, it's designed to be watertight, so to me a leak is a leak and it needs repairing in a proper way and manner".

This matter was dealt with by the experts in section 1.8 of the Joint Note, in which they said this:

The experts are thus agreed that applying a little extra torque to (ie. tightening) the spindle may stop a spindle leak and that, if it does so, this should last until the valve is next operated. This is in accordance with the evidence that Mr Lewis gave in his witness statement, as summarised above, and with Mr Hanson's experience with the hydrant in Wolsey Close. I accept this evidence.

I have already explained (see paragraph above) what is meant by the "slack angle" referred to in the next part of the Joint Note. If the shaft of the spindle is not true and there is a leakage path between the gland packing and the shaft, then turning the shaft very slightly may have the effect of closing the gap and stopping the leak. The effect of the experts’ agreement is that this may only have a temporary effect, because the packing may gradually change shape. The relaxation of the packing referred to by the experts was a reference to stress relaxation: this was explained by Mr Whittaker as the situation when after compression an elastomer no longer has the ability to recover its pre-pressurised configuration, or to do so with sufficient energy so as to achieve a proper seal. Unless the elastomer exerts sufficient force against the adjacent surface it will not be able to overcome the water pressure in order to prevent a leak.

I have to confess to some difficulty in understanding how stress relaxation could occur in the situation described in (iii) of the Joint Note so as to permit an apparent restoration of the seal to fail after a period of time. One possibility is that an increase in the mains water pressure could overcome the seal if there is an element of stress relaxation. Indeed, to some extent this may be consistent with the evidence given by Mr Lewis that I have summarised above about the possibility that a variation in the mains water pressure might restart a leak However, whatever may be the true position, that is what has been agreed by the experts and, in the absence of any more cogent evidence, I feel that I must accept it.

The conclusion that I reach from all of this evidence is that some spindle leaks can be remedied by tightening the valve, at least until the hydrant is next operated, and that others can apparently be remedied by rotating the spindle within its slack angle but that the latter measure may prove to be no more than temporary. In addition, there will be a class of spindle leaks were tightening the hydrant merely reduces the leak but does not eliminate it.

In the light of the agreement of the experts, there is no significant risk that, if a leak is remedied by forcing the gate tighter into its seat or by forcing the thrust collar upwards against the underside of the bonnet, there will be any deterioration before the hydrant is next operated. The fact that there was a leak before this was done does not in my view affect the need for repairs. That need would arise, if it is to arise at all, when the hydrant is next operated and the leak reappears and then cannot be stopped.

It is, in my view, only the last category of leaks, namely those where the leak is apparently sealed by rotating the spindle through its slack angle, that can fall within Mr Herberg's argument that there could be (or is) a real or significant risk that the subsequent minor leaks could worsen within two years to an extent that would be unacceptable, either for fire fighting or for safety reasons. For the reasons that I have already given I do not consider that Three Valleys has shown that this risk exists as a matter of likelihood so as to justify immediate repair on ordinary commercial grounds, unless of course that is what the Act requires.

As will be apparent, the significant difference between a leak through the spindle and a leak through the valve into the outlet chamber or pipework is that the latter presents a risk that the water will accumulate in the outlet so that in the event of sudden depressurisation of the mains upstream of the valve the water lying in the outlet could be sucked back into the mains system.

I should say a little more at this point about the frost or drain valve that is fitted to the outlet of every Type 1 and Type 2 hydrant. These come in various different forms, but essentially each allows water to drain from the outlet chamber or pipework if there is a sufficient head of water above the drain inlet to force the water by gravity through the frost valve and into the pit. Since the frost valves typically have a very narrow aperture the rate at which they will evacuate water from the outlet chamber or pipework will be low. Although I have heard no evidence about it, I would expect that that such a low level of leakage into the pit would be removed either by seepage into the ground or by evaporation. In other words, if there is a properly functioning frost valve and its outlet is not blocked by silt or other debris that has accumulated in the pit, there should not be any substantial quantity of water in the outlet chamber or pipework for any sustained period. However, there will nearly always be a small quantity of water in the outlet that the frost or drain valve cannot evacuate, so that the possibility of some water being sucked back through the valve in the event of a depressurisation of the mains can never be completely eliminated. But in these circumstances the contamination risk would in my judgment be reduced to a level that would be minimal.

However, the risk would not be minimal if either the frost valve was not working or had become obstructed, or if the rate of the flow of water past the main valve was much greater than the frost drain could discharge. In this situation the outlet chamber or pipework would remain full of standing water. Depending on the circumstances this water could well overflow from the outlet and fill the pit. If the water level in the pit was at about the same level as the hydrant outlet, then a fall of heavy rain could probably bring dirty surface water into the pit which could then become mixed with the water in the hydrant outlet - thereby contaminating it.

The Act does not contain any definition of and its provisions do not throw any light on what is meant by "good working order". It is, in one sense, the case for each party that what is meant by "good working order" must be determined by reference to the purpose for which they hydrant is provided. That much seems to me clear from the use of the word “working”, which distinguishes function from, say, appearance. However, the parties take diametrically opposed positions about the nature and extent of that function. As is evident from what I have already said, London Fire takes the position that the only function of a hydrant is to be fit for fire fighting purposes. Three Valleys submits that, in addition to this, a hydrant is a form of valve whose function is not only to permit the supply of water but also to be able to shut it off. Therefore its purpose, says Three Valleys, is to allow the supply of water on demand, and not otherwise: a hydrant that leaks does not do this. Thus its primary case is that a leaking fire hydrant is not in good working order, no matter how slight the leak. I referred to this during the hearing, perhaps a little unfairly, as a "zero tolerance" case.

By way of a secondary case, Three Valleys makes a number of further submissions based on a fire hydrant’s function. It submits that it is part of a fire hydrant’s function that it must not constitute an actual or potential danger, either to users or to members of the public. It subdivides this requirement into hazardous situations which the hydrant might create if not in good working order such as, for example, a leak which overflows from the pit onto the road or pavement and might freeze, and the static condition of the installation - for example, a damaged frame or cover which is in a condition that presents a hazard. As I have already indicated, London Fire does not accept this as part of Three Valleys’ duty under section 57(3) of the Act, but accepts - as a matter of practice – liability for the costs of repairs in these situations.

Three Valleys submits also that if the hydrant presents a contamination risk to the mains water supply, then that also constitutes a failure to fulfil its function. In this context Three Valleys submits that any leak, and certainly any visible leak, presents a contamination risk so that it is part of Three Valleys’ duty to eliminate it.

Finally, Three Valleys submits that even if a visible leak does not present a risk of contamination, nevertheless any visible leak means that a hydrant needs repair in order to keep it in good working order. As I understood the submission, this is based on the premise that a minor leak can develop into a more serious leak to the extent that there will present a significant risk of the hydrant becoming unfit for firefighting before it is next inspected. I have already discussed this argument earlier in this Judgment.

Faced with these conflicting submissions, I must start by considering the purpose of the section. The object, clearly, is to make provision for access to mains water by fire fighting bodies in order to enable them to fight fires. It follows, therefore, that the hydrant itself must be capable of fulfilling this function. In this context, I consider that the “hydrant” referred to in section 57 is not merely the valve and its associated pipework but the whole installation including the frame and cover of the pit. If, for example, the cover was jammed in the frame so that it was not possible to gain access to the hydrant, the hydrant would clearly not be fit for the purpose of fighting a fire, and therefore in my judgment not in good working order. Similarly, if the valve was jammed in the closed position.

Amongst other things the Act provides that it is the duty of a water undertaker, when supplying water to any premises for domestic purposes, to supply only water which is wholesome at the time of supply (see section 68(1) of the Act, as originally passed). Fire hydrants are generally attached to water mains which supply domestic properties and so it follows that they must be constructed, installed and maintained such that they do not present a risk to the integrity of the water supply. Therefore it seems to me that this must be another primary function of a fire hydrant.

In addition, fire hydrants are typically located in pavements or footpaths. It must go without saying that they should not present a hazard to members of the public using that pavement or footpath. Further, I consider that it is self evident that a hydrant should not be in such a condition as to present a risk of harm to those who have to use it. Whatever the precise meaning of "good working order", I would have the greatest difficulty in accepting that a fire hydrant that was in a condition such as to present a source of danger, whether to members of the public or to its potential users, was in good working order.

I therefore consider that a fire hydrant must fulfil each of these functions. That is to say that it must be kept in a condition such that (a) it is fit for the purpose of firefighting, (b) does not present a risk of contamination to the water supply and (c) is not a source of danger to members of the public or potential users.

In order to be fit for the purpose of fire fighting it is obvious that a hydrant must be capable of being opened and closed so as substantially to permit or prevent the flow of water. I have seen nothing in the evidence that would suggest that a hydrant which only permitted, say, 98% of the water supply to flow through it or, alternatively, when it was opened permitted 2% of that supply to escape would not be fit for firefighting purposes. It is London Fire's case that a minor leak (and I am not suggesting that this might be anything like 2% of the available supply) does not impair the function of the hydrant in fire fighting terms at the time and this is not, I think, disputed. The real question in this case is whether a fire hydrant with a minor leak can be said to be in good working order, either simply because it is there or because of the potential for significant deterioration between routine inspections.

In the context of something being in working order, there are other adjectives that come to mind apart from "good". A machine could be described as in perfect working order, or proper working order or, if different, in satisfactory working order. “Working order” is defined in the Shorter Oxford Dictionary as "the condition in which a machine, system, etc, works satisfactorily or in a specified way". On the basis of this definition it is clear that the adjective "satisfactory" adds nothing. I consider also that the adjective "proper" adds nothing to the definition of "working order". However, when a person speaks of a machine being in "perfect working order", I consider that he or she would be describing a machine that ran flawlessly. A proud owner of a vintage motor car, if describing it as being in “perfect working order”, would be conveying the impression that it was working just as well as it did many decades earlier shortly after it left the factory.

As a matter of ordinary language, I consider that "good" denotes a rather better condition than "satisfactory" but a poorer condition than "perfect". Whether it is closer to “satisfactory” or to “perfect” might be a matter of debate, although I would incline to the former. Accordingly, I consider that the adjective “good” does add something to the expression “working order” and that the word is not simply redundant.

Fire hydrants are designed with a 50 year life, and it is apparent from the evidence in this case that some of the Fire hydrants in London have been in place for nearly double that period of time. Mr Ian Whittaker, the expert instructed on behalf of Three Valleys, said in evidence that this did not mean that every individual component of a hydrant would be expected to last for 50 years without replacement, but that the hydrant had to be designed and manufactured so that components like the gland packing could be maintained; in other words, one would not expect the body of the hydrant to fail but one would envisage that a washer might have to be replaced and the hydrant would have to be constructed so that the washer would be replaceable (Day 7, page 10).

In answer to a question from me, Mr Whittaker then said that this implied that whilst a washer, for example, would be in one condition on initial testing at the time of manufacture it would, with time, deteriorate to a condition at which it required replacement. I would have thought that replacement would be necessary at that point when, if it was not carried out, the hydrant would no longer be in good working order. If that is correct, it must follow that some deterioration in the performance of the washer would not prevent the hydrant from being in good working order, although this would obviously be a matter of degree.

Mr Curtis submitted, correctly in my view, that section 57 must be read as a whole and an important feature was the presence of a criminal sanction in subsection (8) for the failure to keep a hydrant in good working order. This, of course, must be read together with subsection (9) which provides the undertaker with a defence if he can show that he took all reasonable steps and exercised all due diligence to avoid the commission of the offence.

I accept that this is a factor to which I should have regard when construing the section, and that it provides a reason for not taking too strict a view about the meaning of good working order. Bearing this in mind, and in the light of my conclusions as to the functions that a hydrant must fulfil if it is to be in good working order, I can now give my views as to what that means in practical terms.

First, I reject London Fire's submission that it is under no obligation to pay for the costs of repairs to a hydrant that is in a dangerous condition. By this I include the costs of repairs either to the cover or to the frame if these repairs are necessary to remove the danger. In my judgment such repairs would be necessary in order to keep the hydrant in good working order.

Second, if the state of a hydrant is such that water is overflowing the pit as a result of a leak, whether through the spindle or the main valve, then the costs of repairing the leak are costs that are necessary in order to keep the hydrant in good working order. Whether this situation is approached as being one in which it is considered necessary to carry out remedial works in order to remove a dangerous situation, or because it is necessary to carry out those works in order to prevent the hydrant from becoming inoperable within the next two years, probably does not matter.

Third, if there is a leak through the main valve at a rate such that it cannot be evacuated through the frost valve with the result that a substantial quantity of water accumulates in the outlet chamber or pipework, with the consequent risk that it might be drawn back through the valve into the mains water supply, the hydrant is not in good working order. I am assuming for this purpose that the frost valve is working and that it is not blocked by silt or other debris.

Fourth, if the pit is filled with silt to a level above the frost valve, with the result that the frost valve will be either obstructed or is likely to become obstructed, or the pit might fill up with water to a level above the opening of the outlet, the hydrant is not in good working order. Here the remedy is to clear out the silt - a task that can readily be performed by London Fire. In my judgment it would not be reasonable for London Fire to incur the cost of having this done by Three Valleys’ contractors (unless they happened to be there to carry out some other work on or in the vicinity of the hydrant). I will turn to some of the practical implications of these conclusions later in this judgment.

Fifth, if the hydrant has a minor leak through the spindle that cannot be remedied by tightening the valve (or by adjustment within the slack angle), or a minor leak through the main valve that is being discharged through the frost outlet, and the extent of the leak is such that it does not cause the pit to fill up with water - in other words, the extent of the leak is such that the water can be removed from the pit either by seepage or by evaporation, or a combination of both - then I consider that such a state of affairs does not mean that the hydrant is not in good working order. Whilst the position should be noted and reported to London Fire, in my judgment immediate repairs would not be required by section 57(3).

However, this last conclusion requires a qualification. It is this. Three Valleys would only be justified in taking no action in respect of such a minor leak if it was understood that, following the receipt of a report of a minor leak, London Fire would take reasonably prompt steps to inspect the hydrant and satisfy itself that it was not likely to become inoperable before it was next due for inspection. It would only be in the knowledge that such a system was in place that Three Valleys would be able to discharge its duty under section 57.

What I have said in the previous paragraphs is intended to provide my answer to Issue A. Unfortunately, I do not think that the issue lends itself to any more concise answer.

It is Three Valleys’ case that a hydrant which is missing a false spindle cap is not in good working order. London Fire disagrees. It submits that, since a variety of false spindle caps is kept on every fire appliance, it does not matter whether a hydrant is fitted with a spindle cap or not. Further, London Fire submits that since there appears to be a market in the scrap metal industry for false spindle caps, it would be pointless to replace them every time they are stolen.

I must confess that at first I was not very impressed with this argument. But as the case developed, I began to see the force of it. If it is the case, and London Fire clearly believes that it is (because it has issued a memo to all Fire Stations about the problem), then it would be pointless to replace every missing spindle cap because one could never be confident that it would still be in place when the hydrant was needed. Far better to keep a supply of false spindle caps on board fire appliances so that there is always one available in case the cap is missing on a hydrant that the Fire Brigade needs to use.

This is probably also an economical solution, as well as a practical one. The cost of sending a maintenance team out to replace a missing spindle cap would be disproportionate to the cost of the cap itself, which I was told was less than £10. At present, it seems that Three Valleys charges about £275 to remove a jammed spindle cap and replace it, and - as far as I can tell from the schedule of charges - it makes no difference whether the Three Valleys has to remove a jammed spindle cap and then replace it, or simply has to replace a missing cap.

Accordingly, on this issue I prefer the arguments of London Fire.

This issue arises out of section 57(5A) of the Act which, for convenience, I set out again below:

The question posed by the preliminary issue is as follows: is the burden on the Defendant to prove that the damage was caused by authorised third party use? Or is the burden on the Claimant to prove that the damage was not so caused?

Both sides advanced arguments in support of their construction of the section based on practical considerations. London Fire, in particular, submitted that if the burden of proof was on London Fire, it would find it very difficult to discharge so that in practice the section would be rendered empty of content.

I consider that these arguments were overstated. First, it is helpful to see with what precisely the section is concerned. It arises only when a hydrant is damaged, so it does not apply where the hydrant has suffered deterioration from normal wear and tear. Second, it has to be damaged as a result of use made of it with the authority of Three Valleys. So it does not arise where there has been vandalism: in that situation, the risk rests with London Fire. Bearing in mind these considerations I suspect that in practice the number of cases where a hydrant is damaged by an authorised user will be comparatively rare (particularly in the light of Three Valleys’ current policy in relation to third party use).

I consider also that in most cases London Fire should be able to tell on inspection whether damage has been caused by vandalism or by operational misuse. It is only if London Fire concludes that the latter is either likely or possible that the section may come into play. London Fire will, or should, also know whether or not its operatives have used the hydrant since it was last inspected. If London Fire had used the hydrant since the last inspection, then it would be likely that any damage caused by operational misuse would have occurred on that occasion. However, that is a matter that London Fire can investigate internally by making appropriate enquiries.

It seems to me that it will be only in cases where London Fire has not used the hydrant since it was last inspected, and where the damage appears to result from operational misuse (as opposed to vandalism), that the question of misuse by an authorised third party will arise. Where these factors are present, I would have thought that it would be more likely than not that the damage will have been caused by someone authorised by Three Valleys.

In these circumstances it would not be inappropriate for London Fire to resist any claim for the cost of the repairs on the grounds that they are covered by section 5A. The onus would then be on Three Valleys to demonstrate that they had not issued a licence to a third party which covered, or might have covered, the hydrant in question at the time when the damage probably occurred.

These considerations lead me to conclude that it would not be unduly onerous for London Fire to have the burden of proving the application of the section. There is, therefore, no ground for seeking to find a way round the usual rule that "he who asserts must prove" (I paraphrase rather crudely the general rule as it is set out in Phipson on Evidence, at paragraph 6-07).

As I have already indicated, if London Fire can establish that a hydrant has been damaged as a result of operational misuse, and that its operatives did not use the hydrant at the material time, it will have discharged - at least on a provisional basis - the burden of proving the application of the section. It will then be for Three Valleys, if it can, to discharge the evidential onus of proving the contrary.

So for these reasons my answer to preliminary issue C is that the burden is on London Fire to prove that the damage to a hydrant was caused by authorised third party use if that is what London Fire believes to have been the case. Whilst the reasons I have set out above do not form part of my answer to the preliminary issue, I hope that they will provide guidance to the parties as to how the burden of proof will arise in practice and how it should be discharged.

It is clear from the terms of section 57 that Three Valleys is not entitled to charge London Fire with the value of the work carried out, but only with the costs that it has reasonably and properly incurred in carrying out work that is required to keep a hydrant in good working order. I have set out earlier in this Judgment what sort of work would fall into that category.

Until April 2005, Three Valleys’ charges for repacking a hydrant, including easing the spindle, were divided into two components. The first was the contractors’ charge, both for labour and materials, and the second was described as "overhead". In his witness statement Mr Ross (at B/185) described the overhead charge in the following terms:

“The overhead charge covers those activities in the process that are not separately charged. These include the administration necessary to receive calls from the public or orders from the brigade, further liaison with the brigade and invoice for the work. Also included in this category is the attendance by CST's, who travel to site to inspect the faulty hydrant, raise the works order and also facilitate the work where required by delivering customer notices in advance for supply interruption and operating network valves.”

In April 2005, Three Valleys’ charges for repacking a hydrant, including easing the spindle, included a further charge. This was a contractor's management fee. Mr Ross described in this as "a cost not included in the unit rates for the job” (B/185).

In April 2006, Three Valleys’ charges for repacking a hydrant included a further charge described as “DLO". Although this had a unit rate of £76, only 50% of it was charged. I suspect that DLO may stand for Direct Labour Organisation and, as I understood Mr Ross (Day 4, at 57), this reflects the overhead in respect of work carried out by Three Valleys itself (because not all work was contracted out, Three Valleys carried out some repairs using its own employees).

Mr Ross said in his witness statement that the charges do not contain any profit element. On its face, this I would accept. However, this assumes that the overhead (and the DLO charge) genuinely reflects an appropriate proportion of operating and administrative costs so that no part of it could be described as profit.

I am unable to reach a clear conclusion about this. Mr Ross was able to assist to a limited extent only when giving evidence because he fairly admitted that the overhead figures were compiled by others in the finance department, who then gave them to him. His task was to allocate the total overhead figure between the different jobs using a forecast of the likely number of the jobs for the forthcoming year. He agreed that Three Valleys did not attempt to compare the actual out-turn with the forecast at the end of each year, but he was adamant that Three Valleys never recovered all of its costs (Day 4, at 61-62).

It was the practice for Three Valleys in about January of each year to issue their proposed charges for the forthcoming year, commencing on 1 April, and so London Fire, and presumably other customers, had an opportunity to make any representations on the charges for that next year. The correspondence contained examples of such queries being raised and dealt with (see, for example, the exchange of e-mails between Mr Hanson and Mr Ross on 14 January 2003, at F/1400, and the subsequent correspondence).

I was told also that the charges had to be approved by Ofwat, and that there are letters in the bundle that show that this happened (see, for example, Ofwat's letter dated 3 February 2004 approving the charges for the 2004/2005 year). However, I suspect that Ofwat might have been more concerned with the charges for water supply and sewerage, rather than the charges relating to the installation and maintenance of hydrants. The procedure for the advance notification of the proposed charges for the forthcoming year was also referred to in Appendix 10 to the MoU.

I accept that when putting the charges together Three Valleys did not intend to include anything but genuine expenditure. I accept also that they did not contain any costs attributable to the Three Valleys leakage teams, as Mr Ross confirmed (Day 4, at 43). But on the material that is before me I am not in a position to conclude that the charges raised by Three Valleys in relation to the hydrants contained the right level of overhead and consequently did not include any element of profit. Conversely, there is no evidence whatever that indicates that the charges did contain an element of profit. The only positive evidence that I have is the assertion by Mr Ross that Three Valleys did not recover all of its costs.

In the light of this evidence and, since London Fire had an opportunity every year to raise any queries on the charges proposed for the next year (and, so far as I can tell, has never questioned whether or not the charges contained any element of profit), it seems to me that it would be wrong for me to find that any part of the charges the subject of this claim included anything but genuine expenditure. In addition, this was not a point taken specifically in the Defence beyond a plea putting Three Valleys to strict proof of the expenses incurred in any repairs.

I must make it clear that in reaching these conclusions I am expressing no view about Three Valleys’ current charges.

In this section of this Judgment I will apply the principles set out above to each of the sample claims. I will also consider the reasonableness of the costs of repair. I have already expressed my conclusion on the question of the possible inclusion of profit or non-attributable overheads in the charges claimed by Three Valleys. I therefore say no more about it.

My findings in relation to each of the sample claims are therefore intended to provide the balance of my answers to Issues B and D.

Job No: 1953658 – 316 Allenby Road, Southall. Repack

The work carried out was a repack, including easing the spindle. The sum claimed is £335, plus VAT, making £393.63 in all. This was in line with the schedule of charges for the year commencing 1 April 2007 (F/1640). This represented the 2006/07 charge increased by 7.5%.

The history of this job can be summarised as follows:

On the basis of Three Valleys’ documents there is no reason to doubt that this old Type 1 hydrant had its spindle gland packed in August 2007. However, there is no evidence to indicate the extent of the leak at that time. The reference in the WMIS sheet to “leakage – detected”, and the fact that repair deadline date was 28 August 2007, tends to suggest a minor leak.

On the face of it, therefore, this looks like a minor leak that would not ordinarily have required repair at that time to keep the hydrant in good working order, with the result that London Fire would not be liable for the cost of the repack. However, I have already held that Three Valleys would discharge its duty in relation to a minor leak if it reported the leak to London fire and could rely on London Fire to inspect the hydrant and form a view as to its condition. But by this time the parties were in dispute about the repairs and Three Valleys might not have known whether London Fire would inspect the hydrant to make an assessment for itself as to the seriousness of the leak. I have therefore considered whether in the circumstances Three Valleys would have been justified in assuming that London Fire would not inspect the hydrant so that it would have been entitled to go ahead and carry out the repairs in any event. However, no arguments were addressed to me on this point and in these circumstances I consider that it would not be appropriate for me to reach any conclusion on it.

Since I have concluded that this was probably a minor leak in the sense that I have described it earlier in this Judgment, I consider that London Fire is not liable for the costs of the repack that was carried out in August 2007. Had I concluded otherwise, I would have accepted that the £393.63 claimed was a reasonable and proper cost.

I have considered whether the fact that Dr Sanders found leakage at the spindle in September 2009 means that the repack was not properly carried out in 2007. I do not think it does: it may be that the cause of the leak was stress relaxation of the gland packing that could have been rectified by tightening the gland bolts.

Job No: 1456370 – 204 Carlton Ave West, Sudbury. Rewasher

The work carried out was a rewasher. The sum claimed is £148, plus VAT, making £173.90 in all. If this included reinstatement of the verge only, then it was in line with the schedule of charges for the year commencing 1 April 2003 (F/1555). The events giving rise to this repair are set out below:

On the basis of the material in the documents, namely that water was "pouring out" and that the job was given a next day priority, I conclude that this hydrant was not in good working order on 27 August 2003 and required repair. It is most improbable that the hydrant would have been left in this condition for several months, and the documents suggest that the valve may have been tightened, and the leak temporarily halted or reduced, on 29 August 2003, and I so find.

London Fire is therefore liable for the costs of this repair. I accept that the £173.90 claimed is a reasonable and proper cost.

Job No: 1435264 – 54 Dudley Rd, S Harrow - Replace hydrant

The work carried out was a replacement. The sum claimed is £570, plus VAT, making £669.75 in all. If this included reinstatement to a path, then it was in line with the schedule of charges for the year commencing 1 April 2003 (F/1554). The events giving rise to this repair are set out below:

This is another case where the documents show that the hydrant was not in good working order, probably as a result of vandalism, because water was flowing down the road. It appears that the work was properly carried out because it was inspected by London Fire a few days later and pronounced satisfactory.

London Fire is therefore liable for the costs of this repair. I accept that the £669.75 claimed is a reasonable and proper cost.

Job No: 1491064 – 7 Northolm, Edgeware - Replace hydrant

The work carried out was a replacement. The sum claimed is £414, plus VAT, making £486.45 in all. If this included reinstatement to a path, then it was in line with the schedule of charges for the year commencing 1 April 2004 (F/1579). The events giving rise to this repair are set out below:

This is an odd case. London Fire’s records show that work to an earlier spindle leak had been repaired the day before Three Valleys raised the report dated 25 November 2003, with a requested completion date of 2 January 2004. London Fire then inspected the hydrant about two months later (I cannot say how promptly, because there is no record of when the defect was reported to London Fire) and reported a spindle leak but described the condition of the hydrant as satisfactory. The work was then completed by Three Valleys about 3 months later.

However, the experts have agreed in their Joint Note, in section 7.4, that replacement of the hydrant was probably appropriate because the necessary spares were not available. By contrast, London Fire in fact inspected the hydrant within a couple of months, perhaps in response to a notification from Three Valleys, and noted that it was satisfactory.

Assuming that it was correctly reported that the glands of this hydrant were broken, and there seems to be no reason to doubt it, then I consider that it cannot have been in good working order at the time when the work was carried out in April 2004 and, probably, at the time when Three Valleys raised its report on 25 November 2003.

On balance I consider that London Fire is liable for the costs of replacing this hydrant. I accept that the £486.45 claimed is a reasonable and proper cost.

Job No: 1678350 – 161 Eastcote Road, Ruislip

This claim has now been abandoned.

Job No: 1361171 – 57 Station Road, West Drayton

The work carried out was renewal of the frame and cover. The sum claimed is £119, plus VAT, making £139.83 in all. The events giving rise to this repair are set out below:

There seems to be no doubt that this hydrant was missing a cover and that a cover was subsequently provided. The problem is that it was the wrong cover, because it was marked with a “W”, instead of “FH”, as Mr Ross confirmed in evidence (Day 4, at 46). That means it was a cover for a washout and not for a fire hydrant.

On this basis I consider that Three Valleys has failed to show that it was entitled to be paid for replacing this cover at the time of issuing its invoice. However, it is clear that this hydrant was not in good working order because it was missing a cover, so if and when Three Valleys fits the correct type of cover it will be entitled to recover the costs of doing so.

Job No: 1837469 – 45c Windmill Lane, Greenford

The work carried out was the supply and fitting of a spindle cap. The sum claimed is £290, plus VAT, making £340.75 in all. The events giving rise to this repair are set out below:

The sum claimed, £340.75, for the simple replacement of a spindle cap, is plainly excessive.

Since I have concluded that a hydrant without a spindle cap is in good working order, this claim must fail. But even if it was justified in principle, Three Valleys would have failed to prove that the sum claimed represented the reasonable costs of doing this work.

Job No: 1843675 – 90 Windmill Lane, Greenford - Repack

The work carried out was a repack, including easing the spindle. The sum claimed is £312, plus VAT, making £366.60 in all. This was in line with the schedule of charges for the year commencing 1 April 2006 (F/1624). The events giving rise to this repair are set out below:

It is clear from the documents that there must have been some form of leak at this hydrant on 5 May 2006. However, the leak was minor and could be stopped by closing the hydrant (see Mr Hanson's witness statement, at paragraph 46, B/121). On the documents it is not possible to say whether the leak had reappeared by the time that Three Valleys came to carry out the repairs on 13 September 2006, or whether a repack was carried out in any event (and irrespective of the existence of a leak) because a work request had been raised.

Given the state of the evidence in relation to the condition of this hydrant, I find that Three Valleys has failed to prove that the work carried out was necessary in order to keep the hydrant in good working order.

Job No: 1847943 – 77 Longmore Avenue, East Barnet - Repack

The work carried out was a repack, including easing the spindle. The sum claimed is £312, plus VAT, making £366.60 in all. This was in line with the schedule of charges for the year commencing 1 April 2006 (F/1624). The events giving rise to this repair are set out below:

On the face of it, therefore, this looks like another case of a minor leak that would not ordinarily have required repair at that time to keep the hydrant in good working order, with the result that London Fire would not be liable for the cost of the repack. However, I have already held that Three Valleys would discharge its duty in relation to a minor leak if it reported the leak to London fire and could rely on London Fire to inspect the hydrant and form a view as to its condition. But by this time the parties were in dispute about the repairs and Three Valleys might not have known whether London Fire would inspect the hydrant to make an assessment for itself as to the seriousness of the leak. As in the case of Claim A, I have therefore considered whether in the circumstances Three Valleys would have been justified in assuming that London Fire would not inspect the hydrant so that it would have been entitled to go ahead and carry out the repairs in any event. However, for the same reason I consider that it would not be appropriate for me to reach any conclusion on it.

Since I have concluded that this was probably a minor leak in the sense that I have described it earlier in this Judgment, I consider that London Fire is not liable for the costs of the repack that was carried out in August 2007. Had I concluded otherwise, I would have accepted that the £366.60 claimed was a reasonable and proper cost.

Job No: 1850063 – 32 The Ridgeway - Repack

The work carried out was a repack, including easing the spindle. The sum claimed is £312, plus VAT, making £366.60 in all. This was in line with the schedule of charges for the year commencing 1 April 2006 (F/1624). The events giving rise to this repair are set out below:

It seems unlikely that the condition of the hydrant would have changed between Mr Hanson’s inspection on 4 October 2006 and the date when the repairs were carried out, 14 November 2006. Since it appears that the pit was pumped out by the repair team, it is likely that they found some form of leak from the spindle since they decided that it required repacking. However, this leak cannot have been severe because otherwise Mr Hanson would have noticed it on his inspection on 4 October 2006. However, I am prepared to accept that Mr Hanson may have overlooked a minor leak at the spindle, particularly if he might have been expecting to find a leak through the main valve.

Somewhat unhelpfully, Mr Hanson decided not to notify Three Valleys of his findings following his inspection on 4 October 2006. He said in cross examination (Day 6, at 21) that he did not want to set a precedent that might encourage Three Valleys to think that London Fire would inspect every hydrant for which they had received a request for a work order. I have to say that I found this to be an unsatisfactory response. Unfortunately, it was symptomatic of the entrenched positions that by this time each party had adopted.

All other things being equal, on the rather confused state of the evidence this might have been another case of a minor leak that did not prevent the hydrant from being in good working order with the result that Three Valleys would not be entitled to recover the costs of the repack. However, it seems to me that on this occasion London Fire has only itself to blame for the fact that the repairs were carried out. Had it notified Three Valleys that the hydrant had been inspected and was not found to be leaking, I suspect that the work would not have been carried out (as was the case at 15 Wolsey Close, discussed at paragraph 71 above) - if it had, Three Valleys would have done it at its own risk.

Therefore, in the unusual circumstances of this particular case, I consider that Three Valleys is entitled to recover the costs of repacking this hydrant. I accept that the £366.60 claimed is a reasonable and proper cost.

Job No: 1866333 – 4 Kingsmead Drive - Repack

The work carried out was a repack, including easing the spindle. The sum claimed is £312, plus VAT, making £366.60 in all. This was in line with the schedule of charges for the year commencing 1 April 2006 (F/1624). The events giving rise to this repair are set out below:

The experts concluded (Joint Note, section 7.1) that the cause of the gusher seen in 2009 was likely owing to the fact that the repack was not carried out properly. Mr Hanson accepted in cross-examination (Day 6, at 29) that it would probably have been impossible to see a leak unless the water was actually bubbling up through the silt. On any view, from the photo at D/520 it would have been very difficult to see any minor leak.

I consider that a hydrant whose pit is filled with sludge up to a level just below the outlet, in which the water can be seen clearly about an inch or two below the level of the rim, is not in good working order. It would have taken very little rain to fill the pit so that the water overtopped the outlet, and thus contaminating the outlet with dirty water.

This is an unusual case. It appears that a repack was carried out on 29 November 2006 but that it may not have been done properly, although I assume that the repair team would have cleaned out the pit at the same time.

In these circumstances I consider that Three Valleys is entitled to recover the costs of attending to the hydrant and cleaning out the pit, but is not entitled to recover the costs of the repack because it was not properly done. I assess these costs at £183.68 plus VAT, making £216.20 in all (F/1618 and 1625).

In reaching this conclusion I am not suggesting that in every case where Three Valleys finds that a pit is full of sludge it would be reasonable for them to send out a repair team; on the contrary, I consider that Three Valleys would discharge its statutory duty by notifying the London Fire of the state of the pit and asking that it be cleaned out.

In summary, my answers to the preliminary issues are as follows:

I am conscious that I have not dealt with every point that has been raised. In order to keep this judgment to a manageable length I have confined myself to the evidence and arguments that I consider bear directly on the issues that I was asked to determine.

I am very grateful to counsel for their very able arguments and to the solicitors on both sides for the very helpful presentation of the documents.