|
|
[Home]
[Databases]
[WorldLII]
[Search]
[Feedback]
South Africa: Supreme Court of Appeal |
[Database Search] [Name Search] [Recent Decisions] [Noteup] [Help]
CASE NO356/98
IN THE SUPREME COURT OF APPEAL OF SOUTH AFRICA
In the matter between
Daniël Mostert
Appellant
and
Cape Town City Council
Respondent
BEFORE: SCHUTZ, SCOTT JJA and MTHIYANE AJA
HEARD: 21 August 2000
DELIVERED: 7 September 2000
W P SCHUTZ
Negligence - burst water main - cause uncertain - inappropriateness
of res ipsa reasoning - extent of municipality’s duty to prevent
damage to property - insurance - pressure testing - replacement - cost
of
compared with damage done - wrongfulness apparent if negligence proved - which
it was not - new evidence
refused.
________________________________________________________________
J U D G M E N T
________________________________________________________________
SCHUTZ
JA:
[1] The key issue is whether the appellant (“Mostert”) has proved that the
respondent, the Cape Town City Council (“the
Council”), was negligent in regard to its 62 km water pipeline which
leads
from the Steenbras reservoir to the Molteno reservoir situated above the
city. The main negligence relied upon is either not replacing
large parts of
the pipeline, or not regularly testing it under pressure. An alternative
contention is that the Council was negligent
in not insuring itself against
claims by property owners. The careful but succinct judgment of Tebbutt J, a
quo, makes it unnecessary for me to provide the parties with a repetition of
the details of the evidence. I shall confine myself to
an
outline.
[2] This two foot six inch pipeline (762 mm) is made up of some
17600
cast-iron pipes, each weighing about two imperial tons, joined by
spigot and socket joints sealed with lead. Their thickness ranges
from 27/32"
(21.4mm) to 1 3/16" (30.2mm), depending on water pressure from place to place.
For almost its entire length the pipeline
is buried underground at various
depths. Originally it was used as a trunk line delivering “raw”
water to the Molteno
reservoir. Later it was altered into a distribution line,
with smaller lines leading treated water out of it along its length. Some
37 of
its 62 kilometres lie under roadways in built-up areas. Much of it runs under
Voortrekker Road, a main arterial thoroughfare
bearing traffic through the
northern suburbs of Cape Town from Belville towards Cape Town, and passing
through the suburbs of Parow
and Goodwood. Voortrekker Road carries large
volumes of traffic, including many buses, trucks and other heavy vehicles. The
significance
of load imposed on the pipeline was a much debated question in the
trial. The fact that the line runs through built-up areas means
that, depending
upon the topography along its length and the ability of the stormwater drains to
carry off water, a burst in the
main may lead to substantial damage to
property.
[3] In the 1970s the state of the pipe was considered such that it
was worthy of being coated with an internal cement lining in order
to limit
further corrosion. Cement lining is an expensive procedure which is not
undertaken unless a pipeline is regarded as basically
sound. This lining was
done between 1974 and 1982. The process involved cutting open 150m sections and
cleaning out internal corrosion
which had accumulated over the previous more
than 50 years, with a scraping device pulled through the pipeline. This process
may
have damaged the pipeline at isolated spots. After each scraping the 150m
section was inspected visually for soundness and then
the cement lining was
applied.
[4] In 1981, after a burst in the pipeline, Professor Ball, a
metallurgist, was asked to examine specimens in order to determine whether
corrosion had reached the point where the pipeline was nearing the end of its
useful lifetime. His conclusion was that the metallurgical
condition of the
cast-iron was good and did not account for the failure. Some extraneous impact
or soil movement must have been
the cause of the cracking. Records pertinent to
this test are largely non-existent and the details are lost. But what is clear
is that the Council officials were concerned about the burst, considered whether
the pipes had corroded to an undue extent, approached
an expert to examine the
question and obtained an answer favourable as to the state of the pipes.
Professor Ball explained that
the object of his tests was to eliminate corrosion
as a cause of fracture. He was not asked to conduct a wider test to ascertain
whether the pipeline was suspect in some other respect, or to enquire into what
may have caused the extraneous impacts or soil movements.
Mr du Plessis a
consulting engineer called on behalf of the Council readily conceded that a
wider investigation would have been
desirable, but he was not asked what should
have been investigated. Whether any such investigation would have produced
anything
of value is a matter for doubt, particularly in the light of the
inconclusivity of investigations made in the 1990's before the trial.
What
emerges is that the Council’s employees were well aware of how extraneous
impacts or soil disturbances may have been
caused, but this awareness was of no
real help in finding solutions. In the result Mostert, as plaintiff, has failed
to establish
that further investigation in 1981 would have led to the prevention
of later bursts.
[5] On 18 December 1990 there occurred a burst that has led
to this case. It took place under Voortrekker Road in Parow. Extensive
flooding of neighbouring premises resulted, including Mostert’s furniture
shop, Furni-Scene. He claims that damage to stock,
for which he holds the
Council liable, amounted to R307 458. Quantum and liability were separated at
the trial and the former stood
over. Questions arise as at the cause of the
burst, its foreseeability and the extent of the Council’s duty to have
taken
steps to prevent it or to compensate loss consequent upon it.
[6] The
stand of the witnesses for the Council is that, although the pipeline has
suffered some corrosion since it was inaugurated
as far back as 1921, so that
the metal wall has been somewhat thinned, it is still sound, still serves its
purpose, and will continue
to do so for many years to come. In the words of one
Coetzee, who was until his retirement senior superintendent of reservoirs and
trunk mains, “it’s a great pipe.” Professor Ball examined
further specimens from the pipeline in 1994 and concluded
that “it is a
very good water main which is performing well.” Coetzee’s training
was as a fitter and turner.
Over his 40 years experience with the pipeline he
has always found it to be of sound, high quality cast-iron. The Council’s
case is that the bursts over the years for which no definite cause has been
determined, are attributable either to randomly spaced
imperfections in the
cast-iron inherent in its nature, which lead to failures at unpredictable places
and intervals; or to some force
applied to a pipe as a result of a variety of
possible events, some of which will be mentioned later. Mostert’s case
on the
other hand, (not always fully supported by his witnesses, as will be
seen) is that the pipeline has reached the end of its useful
and risk-free life,
so that either it must be replaced, or some additional precautions must be taken
to avoid damage to property
beside the road. The initial list of suggested
precautions was a long one. It shortened as the case proceeded. Broadly
speaking,
the argument is that corrosion and load stresses imposed by
contemporary heavy traffic (heavy in both senses), when compared with
those of
1921, have made the pipeline as a whole sub-standard (or at least those parts
where it lies under a road).
Statistics
[7] A main argument
advanced by two of Mostert’s witnesses needs to be dealt with at the
outset. It has to do with statistical
probability. It was advanced both by
Professor Rooseboom, a civil engineer, and Professor van Rooyen, a metallurgist
specialising
in metal fatigue. Neither is a statistician. In the appeal it was
abandoned by Mr Smit SC, for Mostert, but was persisted in by
his junior, Mr van
Rooyen, in reply. Whereas the pipeline is 62 km long, it was statistically
significant, said the professors,
that three unexplained bursts occurred over a
period of four years (spanning 1979, 1981, and 1983), on a stretch of 11 km,
where
the pipeline lies beneath or beside Voortrekker Road, under tar. Using a
“moving average”, Professor Rooseboom contended
that, as these were
the only unexplained bursts during the four years 1979 to 1983, a non-random
pattern was revealed. This needed
explanation. When one looked for an
explanation, it was to hand. All these bursts occurred under tar. Ergo, the
villain was traffic.
I have somewhat simplified the argument, but this is what
it came to. The conclusion so reached largely underlay the evidence of
both
professors. Corroboration for their conclusion as to the cause of the bursts
was then sought in a variety of directions. Furthermore
it was contended that,
once the pattern and therefore the probable cause of the bursts had become
apparent by 1983, the Council should
have heeded the statistical alarm bell and
taken appropriate action to avoid further bursts.
[8] The validity of this
reasoning was challenged by Professor Underhill, who established, in my view,
that the facts relied upon
were consistent with random occurrence. As
experience of life and a tenuous grasp of mathematics teaches one, the theory of
probability
does not always match the legitimate expectations of common sense.
Professor Underhill, who is a statistician, was instructed that
four bursts had
occurred for which no cause could be determined, in 1972, 1979, 1981 and 1983.
He used a computer program to generate
four random bursts over a 62 km pipeline
10 000 times, and found that in 17.1% of the trials three bursts were clustered
over a distance
of 11 km over four years. This is about the same chance as
throwing a six with a dice. It is generally accepted, he said, that
for a
chance to be statistically significant, in the sense of indicating that the
result is non-random, it must be below 5% (as compared
with 17.1% in the tests).
Accordingly no significance could be attached to the proximity in space and
time of the three bursts between
1979 and 1983. They could have occurred at
random. The limited nature of this conclusion needs to be stressed. It does
not mean that the bursts were random in their occurrence. It only means
that they could well have been random, so that no statistical support is
afforded to Mostert’s case by the seeming coincidence of the bursts in
space and time. As to Professor Rooseboom’s
method of using a
“moving average”, Professor Underhill regarded this use as
unscientific. It was wrong to have selected
the very area where breaks had
occurred and then apply the theory of probability to it. This was like betting
on a horse race after
it has been run. Statistics may therefore be put on one
side.
[9] An even more simplistic argument put forward by Rooseboom and van
Rooyen, that the pipeline is demonstrated to be sub-standard
because it has
burst without apparent cause on several occasions, may also be put on one side.
The facts yet to be discussed reveal
that there is a variety of possible causes
for these bursts, some of which at least are not indicative of the inadequacy
of the
pipeline as a whole, or certain definable lengths of it.
Notwithstanding this, both professors continued to display a strong attachment
to this simple theory.
Possible Causes
[10] I turn to the question; what may have been the causes of the hitherto
unexplained bursts? There was agreement that failure
may have stemmed from relatively small cracks in the cast-iron. These result
from gas bubbles trapped in the material during the casting process. The
existence of cracks does not in itself point to manufacturer’s
error. It
is, even without error, a characteristic of cast-iron. Professor Ball’s
opinion of the further specimens of pipeline
examined in 1994 was, as in 1981,
that the pipes were in good shape metallurgically. The reason for the bursts
over the years (in
addition to the extraneous impact or soil movement mentioned
in 1981) may have been the presence of a relatively small number of
“active” cracks, that is cracks of a length sufficient such that
they might propagate under the designed working stress,
ultimately leading to
bursts. There is no way of knowing where each “rogue” pipe may be
or whether or when it may fracture.
And even if there were a few batches with
defects due to errors in manufacture, there is no knowing how the individual
pipes were
dispersed over the length of the pipeline. (The pipes were
manufactured in Scotland and then shipped by sea and land during and
after the
First World War. With the various transshipments it cannot be assumed that a
batch would remain united and end up end
to end). According to the professor,
active cracks leading to failures scattered in space and time were a price that
was paid when
choosing cast-iron over steel. Steel is less brittle and less
liable to burst, but on the other hand, it was more expensive than
cast-iron at
the time, and is more subject to corrosion.
[11] Professor van Rooyen went a
discrete step further than Profesor Ball. His view was that what had happened
was that “passive”
cracks (too short to propagate to the point of
fracture under designed loads) had developed into “active” cracks.
The
reason for this was that modern traffic imposed stresses on the pipes which
they were not designed to bear. Loads imposed time and
time again (cyclically)
might deform the pipes recurrently and cause metal fatigue, leading to the
spreading of the cracks. This
process might have been helped on by corrosion.
According to this theory, at least that part of the pipeline which lies under
heavily
trafficked areas is below modern standards. Professor Ball rejected the
notion that the propagation of passive into active cracks
was the cause of the
failures. He remained of the view, despite Professor van Rooyen’s
evidence, that the causes were either
the propagation of active cracks or some
extraneous stress imposed upon the pipes. Otherwise the pipes were quite strong
enough
for the job. Professor van Rooyen claimed that Professor Ball had failed
to take metal fatigue into account. This is not correct.
Metal fatigue was
allowed for in the safety factor to which I shall now refer.
[12] Apart from
examining the state of the metal, Professor Ball concluded that a large safety
factor had been built into the specification
of pipe strengths. It was left to
Mr Ramsay, the Council’s principal engineer for design and planning, to
calculate the overall
actual safety factor in regard to imposed loads. I shall
return to his evidence later. A great deal of evidence was given as to
what
this safety factor really was.
[13] Professor van Rooyen placed reliance upon
a 1957 American specification for laying underground pipelines, which recommend
a safety
factor of 2.5 (ie the pipes should be two and half times as strong as
the strength needed to withstand the assumed stresses). According
to van
Rooyen’s calculations the actual safety factor in the pipeline was
considerably less than 2.5, so that it was “a
really borderline
case.” This calculation involved certain assumptions. An important one
was the nature of the “field
conditions”. By this is meant, the
bedding of the pipe. For instance, a tamped sand bed shaped to fit the pipes
affords
the best support. Also important, as far as traffic loads are
concerned, is the depth at which the pipe is laid. The deeper the
pipe the more
are traffic stresses above it dispersed. Ramsay recalculated the safety factor,
taking into account the actual bedding
conditions (as observed by various
Council witnesses over the years), the actual depths at each burst, and the
reduction in thickness
of the pipe due to corrosion. His conclusion was that
the safety factor was well over the recommended factor of 2.5. In addition
the
actual strength of the metal (as tested) considerably exceeded that assumed in
the specification. This provided a further cushion,
or safety factor. Mr du
Plessis was of a similar view. He was of the opinion that even with the thinned
wall, the safety factor
was 2.6, which meant that metal fatigue was not the
reason for failure.
[14] Professor van Rooyen made an unconvincing attempt to
keep his theory as to the propagation of passive into active cracks alive
after
Mr Ramsay’s demonstration of the actual safety factor had undermined the
professor’s calculations (which had also
included an error). I have
mentioned already that he made the unfounded suggestion that Professor Ball had
ignored metal fatigue.
He then suggested that the 1957 American
specification, on which he had based his evidence, was deficient in that it also
did
not make allowance, or sufficient allowance, for metal fatigue caused by
cyclical traffic stresses (although he once contradicted
himself on this point).
He did not fare well in cross-examination, as the American specification is not
based only on theoretical
calculations, but also upon practical experience,
which includes cases of pipelines under roads. Under pressure in
cross-examination,
at one point the professor fell back upon the proposition,
already mentioned, that the safety factor was proved to be insufficient
simply
because there had been failures. In the result it seems to me that the
specification chosen by Professor van Rooyen himself,
when properly applied
works against his theory of the propagation of passive cracks.
[15] The
principal difference between Professors Rooseboom and van Rooyen on the one
side, and Professor Ball and Messrs Ramsay and
du Plessis on the other, as may
be seen, is that, whereas the latter group sees the principal cause of fractures
to be original pre
- 1921 cracks leading to bursts at unpredictable times and
places; the former group sees the principal cause as traffic stresses
imposed
on a pipeline which is no longer strong enough to bear them. In other words the
design strengths of 1919 were insufficient
for the actual conditions of the
1970' s and 1980's. Reference was made to the bathtub curve. The image is of a
lengthwise section
through a bath. The horizontal axis is time, the vertical
events. The incidence of events is high at the beginning and the end of
the
lifetime of the subject. Thus a person might need the attentions of a doctor
frequently in early childhood and old age, but
less in between. Similarly a
motor car’s need of repairs. Also a pipeline may have many failures after
installation and again
at the end of its lifetime, as corrosion reaches a
terminal level. The Mostert counsel claim that the pipeline is beginning to
mount the final upward curve (as will appear below, the Mostert witnesses are
less adamant). The Council representatives insist
that it is still on the
level plain. Professor Ball’s view is that metallurgically the pipeline
is not at the end of the bathtub
curve. Ramsay is of the view that it was
conservatively designed and installed and is still in good condition. Bad
construction
is a possible cause of failure but there is no evidence of it.
When regard is had to the safety factor, which makes allowance for
traffic
stresses, traffic is not the cause.
[16] Both groups recognize that there
may be other causes of failure, the existence or non-existence of which is
difficult or impossible
to establish. Thus, apart from cracks formed in the
casting process, handling damage may have been caused on the voyage from the
foundry in Scotland to the site, or upon installation on site. As mentioned
already, bad bedding can impose unusual stresses, leading
to failure. As far as
the observations of Council officials who gave evidence go, observations made
when repairing, cement-lining,
deviating or connecting the pipeline over the
years, the original bedding conditions were good. But this does not exclude the
possibility
of subsequent damage. For instance, an undetected leak at a joint
may erode the bedding. The whole pipeline is patrolled once every
two days, but
this does not eliminate this possibility. Similarly damage may be caused to the
bedding by contractors not under the
Council’s direct control working near
the pipeline. An example is the laying of a telephone cable. The pipeline
passes through
the jurisdictions of several other municipalities, which
complicates its protection. This despite the Council’s inspections
and
its rules as to notification of such works, so as to allow its own supervisors
to be present. Another possible cause is a heavy
object falling off a truck.
Certain of the Council officials were of the view that the vigorous scraping of
the pipeline prior to
its lining with cement may have damaged it and been a
cause of bursts. Barnard, senior superintendent of reservoirs and trunk mains
(he is Coetzee’s successor) had a “gut feeling” that this was
the cause of the two Parow bursts (presumably 1981
and 1990). Ramsay suspected
the same in respect of the 1981 and 1983 bursts. (Knock those out and how do
even the van Rooyen
probabilities look?) Steps taken to preserve the pipeline
by lining it may thus have damaged it at isolated places. So that, all
in all,
there are possible causes enough.
[17] Before pursuing the matter of
causation further it is desirable to set out the history of pipe failures. The
Council’s
record of bursts since 1972 is complete, if sometimes lacking in
detail. Before that date the records are sporadic. Known failures
occurred in
1950, 1963, 1964, 1970, 1972, 1974, 1979, 1981, 1982, 1983, 1987 and 1990.
Several of these may be eliminated from the
discussion, because a cause other
than pipe failure has been established, or cannot be excluded. The 1964 failure
was due to defective
couplings. Nothing is known of the 1970 failure other than
the date and place of its occurrence. The 1974 burst was due to damage
caused
by a pipe lining contractor. The 1982 burst was due to a mistake made by the
crew pressuring the main. That leaves 1950,
1963, 1972, 1979, 1981, 1983, 1987
and 1990.
[18] To take these in turn: all that is known of the 1950 failure
is that it occurred under Voortrekker Road. The record of the 1963
failure
shows only that it occurred near Firgrove Station, Firgrove, that it was a pipe
failure, and that the condition of the adjacent
pipe was “good.”
All that is known of the 1972 failure is that it occurred under a broad island
dividing the main road,
i e it was not subject to traffic load. The 1979
failure is interesting. It took place not under Voortrekker road, but under a
tarred
cul-de-sac beside it, called Artreco Road, leading to the Epic factory in
Maitland. The pipe was buried 6.6 feet beneath the surface.
The 1981 failure
took place in Voortrekker Road, off de la Rey Road, in Parow. It was this burst
that led to the submission of
test samples to Professor Ball. The depth of the
pipe is not known. The fracture lay under a major intersection. The 1983
failure
took place at the corner of Voortrekker Road and Loop Street, Maitland,
but not in the intersection. The depth of the pipe was 3
feet. Ramsay offered
it as his view (as already stated) that the burst may have been caused by the
cement liner who had worked
on the section not long before. The 1987 failure
occurred in the veld at Faure some 10-15m from the tarred road. The depth of
the pipe was four feet. Finally, the 1990 burst took place under Voortrekker
road, Parow at a depth of 3.34 feet. Mr Flower is
a professional engineer
employed by the Council. He inspected the site of the burst and found a
fragment of pipe which, because
of differently hued corrosion stains, suggested
that the cause of this failure had been an old crack, about 200 mm long (which
would
qualify it as an “active” crack).
[19] The approximate
distances of these bursts from the Molteno resevoir are:
1983 7km
1979 8"
1990 17"
1981 18"
1972 33"
1987 40"
[20] The
grouping of the 1979, 1981 and 1983 bursts over 11 km may be seen in the first
four items. What should also be observed
is the sometimes clustering of
bursts, as also their wide dispersal over the length of the
pipeline.
[21] Whilst disavowing any statistical approach, it is significant
to notice that some of these failures exhibit features which point
away from
traffic load as their cause. Both the 1972 and 1987 bursts took place under a
surface not used for traffic. Moreover,
the latter one was covered by four feet
of overburden, so that traffic load would largely have been dispersed. The 1979
burst can
also be classed with these two. True it was under tar (for what that
is worth) but beneath a cul-de-sac, where neither fast nor
frequent traffic
would be expected. And it was relatively deeply buried at 6.6 feet.
(Flower’s uncontradicted evidence was
that laying a pipe about a meter
under the surface was standard practice and was used in pipe laying in the
Council’s system
and in many other major pipe systems). It is true that a
learner steam-roller driver might have used the cul-de-sac for practice,
but
that seems unlikely. So that there are three instances where it does not appear
that traffic load was the cause of failure.
What was the cause in these cases?
Another feature is that there is not a regular pattern of bursts occurring under
important intersections,
which bear a particularly heavy traffic load. One must
look not only at the historical bursts, but also at the main intersections
where
bursts did not occur. Du Plessis gives as an example Vanguard Drive,
which carries heavy traffic. There have been no bursts under this
intersection.
[22] For all these reasons I do not think that Mostert has
proved on a balance of probabilities that traffic stresses caused the 1990
burst, even though such stresses are an important suspect. One simply does not
know. I do not, therefore, agree with Tebbut J’s
conclusion that
“on the probabilities the most likely cause of such failures was the
effect of traffic loading on the pipeline
at certain points causing passive
inherent defects in the pipe to be propagated into active defects leading to
pipe fracture.”
Possible Remedial Measures
[23] But I do
agree with Tebbutt J that, assuming his finding as to causation was right,
Mostert failed to show that there were remedial
measures available to the
Council, which in all the circumstances it was under a duty to have taken and
which it did not take, from
1983 onwards.
[24] The suggestions as to what the
Council should have done were quite numerous. During the course of the trial
several of them
were abandoned. They tended to be tentative, both as to exact
method and cost. To an extent this is understandable. It would be
expecting
much of a plaintiff in a case like this to have him fully design a variety of
alternative measures and cost them accurately.
Without laying down any standard
of proof in this regard, I shall approach the matter on the basis that if
Mostert has put forward
a proposal within the realms of “practical
politics”, it may behove the Council to parry the proposal. After all,
the
Council knows the facts about its pipeline, and has the facilities and
indeed the duty to consider such matters, which affect its
ratepayers.
[25] As to the abandoned proposals, I shall do little more than
mention them. Magnetic particle inspection in order to locate cracks
from the
inside of the pipe could not be carried out, because of the intervention of the
cement lining. Encasement of the pipes
in concrete was abandoned, mainly
because the cost would not have been much less than that for replacement. A
proposal to place
concrete slabs over the pipes was also dropped. The internal
lining of the pipeline with PVC piping was not persisted in as a remedy,
mainly
because of the cost and the fact that the water flow would have been reduced
drastically. The installation of an automatic
valve closing system, relied on
at the trial, was not pursued in the appeal.
[26] That left insurance to be
taken out by the Council, regular pressure testing to locate weak spots and
replacement with steel
pipes over the whole or a large part of the 62 km of
pipeline.
Insurance
[27] The engineering witnesses for Mostert expressed the view that in
planning its activities the Council should have had regard to
the cost to
particular ratepayers of activities - such as reticulating and
selling water –
which benefited the authority and the general body of
ratepayers. The cost to
individuals should be shared by all, it was argued.
Whatever the philosophical
merit of this suggestion may be, the matter must
be decided according to the
common law (no resort was had to the
Constitution). The Council cannot
insure unless it has an insurable
interest. It has no such interest unless it is
liable to someone for damage
caused by the escape of its water. It is not liable
unless it has acted
negligently or deliberately. To hold otherwise would be to
impose an
absolute liability on the Council - to make it an insurer. That is a
matter
for the legislature, not the courts. In the result the insurance proposal
lacks a basis in law.
Pressure Testing
[28] The proposal put forward by Professor van Rooyen was that at “regular
intervals” sections of the pipeline should
be subjected to pressures 50% more than normal, in order to seek out defective
pipes
with a view to their replacement.
[29] Mr Ramsay, the only pipeline
expert called on the subject, regarded the proposal as both impractical and
dangerous, dangerous
in that such tests could damage the pipeline. First
impracticality. Because it serves as a distribution line, a test would entail
shutting off the valves in pipes leading out of it, which would lead to affected
areas being without water. Moreover, the seals
provided by valves that would
have to be closed were so imperfect that a proper test was not feasible. As far
as cost was concerned,
testing would have to go on all year in order to complete
one test, as it was not practicable to test more than a kilometre at a
time and
tests would presumably have to be performed over week-ends when there was less
traffic. An additional problem was, at what
intervals should tests be
performed? Van Rooyen was of no real help on this practical question and
Ramsay did not know what the
answer to it was. On top of all this (and this is
the danger aspect) Ramsay was of the view that such tests could have a
deleterious
effect, in that they could cause joints to leak, leading to erosion
of the bedding, with consequent possible damage to pipes. Van
Rooyen had no
effective counter to these objections, and I consider that the court a
quo was correct in rejecting the proposal as a solution.
Replacement
With Steel Piping
[30] The proposal was that the Council should have replaced the whole or at
least a substantial part of the pipeline with steel piping.
This proposal does have the virtue of presenting a practical engineering
solution. The question is whether it should have been adopted, given all the
circumstances, particularly the cost of installation
as compared to the
likelihood and cost of damage should the old pipes remain in
service.
[31] The proposals as to the extent to which the pipeline should be
replaced were imprecise and varied from time to time. In his
two expert’s
summaries Rooseboom did not propose replacement at all. In his evidence he
suggested replacement where the risk
was great, but this only if no other remedy
could be found, or if the Council failed to contrive a means of compensating
affected
owners. What he did not propose was wholesale replacement. Van Rooyen
conceded that he did not know where the pipeline was on the
bathtub curve, and
that replacement was not called for unless one knew that one was on the final
upward curve. No witness unequivocally
stated that the whole pipeline needed
replacement. In argument Mr Smit suggested that the Council should launch an
investigation
to ascertain the built-up areas in which the stormwater system
might not cope with a burst and replace in those areas.
[32] The cost of
replacing the whole was estimated by Ramsay to be R118 000 000 in 1989 terms and
R93 000 000 for the section under
roadways in built-up areas (the cost per
kilometre is higher in such areas). As against this cost, Flower stated that,
except in
1981 and 1990, the stormwater system coped with the bulk of the water.
Only in those years was damage caused to property other than
that of the
Council. The damage in 1981 does not appear to have amounted to a great deal
(possibly of the order of R140 000). As
already mentioned Mostert’s claim
(1990) is for R307 458, although his was not the only property damaged. It is
clear that
the loss so far suffered by property owners adjacent to the pipeline
in built-up areas is far less than even the annual cost of servicing
and paying
off a loan obtained to pay for a new pipeline, never mind the cost of the
pipeline itself.
[33] The classic test for negligence is stated in Kruger
v Coetzee 1966 (2) SA 428 (A) at 430 E-G, as follows:
“For the purposes of liability culpa arises if -
(a) a diligens paterfamilias in the position of the defendant
(i) would foresee the reasonable possibility of his conduct injuring another in his person or property and causing him patrimonial loss; and
(ii) would take reasonable steps to guard against such occurrence; and
(b) the defendant failed to take such steps.”
[34] Requirement
(a)(i) is satisfied. The Council could foresee in 1983 that pipes might burst
again, causing property damage and
consequent patrimonial loss. But as Holmes JA
pointed out (at 430 G):
“Requirement (a)(ii) is sometimes overlooked. Whether a diligens paterfamilias in the position of the person concerned would take any guarding steps at all and, if so, what steps would be reasonable, must always depend upon the particular circumstances of each case.”
[35] It is certainly not the position that the
Council has adopted an attitude of indifference, content that damage might fall
upon
the luckless. Active steps are taken to prevent or minimize damage. I
have referred to the two-daily patrol, aimed at detecting
leaks and activities
near the pipeline which might cause it harm. In addition, repair crews and
engineers are on 24 hour standby,
to deal with bursts as expeditiously as
possible when they should occur. But the question is whether the Council was
obliged to
go further and replace a substantial part of the pipeline. In the
words of Prof J C van der Walt, quoted from Lawsa in Ngubane v South
African Transport Services 1991(1) SA 756 (A) at 776 I:
“There are . . . four basic considerations in each case which influence the reaction of the reasonable man in the situation posing a foreseeable risk of harm to others: (a) the degree or extent of the risk created by the actor’s conduct; (b) the gravity of the possible consequences if the risk of harm materialises; (c) the utility of the actor’s conduct; and (d) the burden of eliminating the risk of harm.”
[36] When
considerations (a) and (b) are weighed against consideration (d), affordability
and proportionality between the loss which
may be suffered by occupiers of land
and the cost of replacement to the Council must be examined: Administrateur ,
Transvaal v van der Merwe 1994(4) SA 347 (A) at 363 C-H (the enquiry in that
case was conducted in the context of wrongfulness, but that does not affect the
principle or its application).
[37] Returning to the facts of this case, I
agree with Tebbutt J’s conclusion that it would not have been
reasonable to have
expected that the Council should have expended either R 118
000 000 or R 93 000 000 in replacing the pipeline between 1983 and 1990.
Nor do
I think that it is to be expected to spend a lesser, but still large sum on
replacing pipes where the risk of flooding is
regarded as greatest (no easy or
certain enterprise to determine where).
[38] Accordingly I do not think that Mostert has established that the Council
should have taken further steps to guard against flooding
losses.
[39] I would add, however, that if Mostert had established that the
pipeline was no longer in the plains but on the final upward slope
of the
bathtub curve, or if in the future it should become apparent to a skilled
observer that that has happened, very different
considerations might apply. In
such a case it might be unreasonable not to accept the cost of
replacement.
Res Ipsa Loquitur
[40] Mostert’s counsel have sought to bolster their client’s case by resort to
the form of reasoning by inference that goes under the label res ipsa loquitur. It is described in Hoffmann and Zeffertt’s The SA Law of Evidence 4 ed at 551 in this way “If an accident happens in a manner which is unexplained but which does not ordinarily occur unless there has been negligence, the court is entitled to infer that it was caused by negligence.” Reliance has been placed on two American decisions, George Foltis Inc v City of New York 21 NYS 2 800 and Adam Hat Stores v Kansas City (M) 316 SW 2d 594. In these cases res ipsa loquitur reasoning was applied. A passage in the George Foltis case was quoted in the Adam Hat case and in the heads filed on behalf of Mostert to the following effect:
“Cast-iron water mains which are properly laid four feet underground ordinarily do not break, any more than ordinary trains are derailed, missiles fly, or elevators or walls fall; and when such a main does break the inference of negligence follows in logical sequence.”
[41] On the
evidence before us cast-iron pipes properly laid do sometimes burst, for the
reason given by Professor Ball - slow propagation
of active cracks - and for a
variety of other reasons not necessarily consistent with negligence on the part
of the pipe’s
owner. Professor van Rooyen conceded that some of the
bursts may have been caused by defects lying dormant in the pipeline since
1921.
When the George Foltis passage was put to Professor Ball he said that he
disagreed with it, because, as he said, cast-iron pipes all over the world do
break,
very occasionally. This evidence is not contradicted. Accordingly
res ipsa reasoning is not helpful in this case, either because its first
requirement is lacking (“an accident . . . which does not ordinarily
occur
unless there has been negligence”); or because, if an initial inference
could be drawn, it has been rebutted by the evidence.
Conclusion On
Negligence
[42] Viewing the evidence as a whole I am not persuaded that
Mostert has proved, what in the final analysis he had to prove, that
the Council
was negligent in failing to take further steps to prevent the burst in
1990.
Wrongfulness
[43] I have approached this case as one raising
questions of negligence, whereas an unbending adherence to logic might dictate
that
wrongfulness is the prior enquiry, and the question of the reasonableness
of expecting the Council to replace the pipeline might
have been dealt with
under that heading. Logic is one thing, utility sometimes another. As was
pointed out by Scott JA in Sea Harvest Corporation (Pty) Ltd and Another v
Duncan Dock Cold Storage (Pty) Ltd and Another 2000 (1) SA 827 (A) at 837 H,
in many if not most delicts the issue of wrongfulness is uncontentious,
as the action is founded upon conduct which, if held to be culpable, would
be prima facie wrongful. This is such a case. If the Council was
negligent in not preventing the 1990 burst I have no doubt that the
community’s
sense of what the law ought to be would demand that liability
be imposed upon the Council (cf The Municipality of Cape Town v Bakkerud
(SCA) unreported 29.5.2000). After all, the Council leads across densely
populated land a large volume of water under pressure,
and then exercises
exclusive control over it. Whatever its contrasted social utility, this is the
equivalent of walking ones tiger
across the forum.
New Evidence
[44] After judgment had been reserved but before it had been delivered, on 25
June 1995, within at most a few hours of each other, three further bursts
occurred under Voortrekker Road, two in Parow and one in Goodwood, In two
of the cases, and probably in the third also, an old crack was
found in the
remaining fragments. Mostert applied before Tebbutt J to have
the Council
reports reflecting these facts accepted as evidence.
Mostert’s attitude then was
that there would be no need to re-open the
trial after such acceptance, but that if
it was considered necessary, that
would have to be done. The Council’s attitude
was and is that if the
reports are admitted the trial will have to be re-opened.
[45] Tebbutt J
had a discretion to admit the evidence (see eg Oosthuizen v Stanley 1938
AD 322 at 333, Mkwanazi v Van der Merwe and Another 1970 (1) SA 609 (A)
at 616 B-617 D). The principles guiding the exercise of such a discretion are
set out in these cases.
[46] I am of the opinion that Tebbutt J was correct
in refusing to admit the further evidence. Although Mostert surmounted the
first
hurdle (why did he not lead the evidence before? - because it did not
exist), he failed to clear some others. First, as these bursts
occurred in 1995
they fail to throw any light on one of the main issues - what should the Council
have foreseen before 1990? Secondly,
in two instances, and probably in all
three, the state of the fractures appears to be consistent with the
Council’s case,
that the probable cause of past failures was the
propagation of old active cracks. The evidence does not, therefore, promise to
alter the result of the trial. Thirdly, there is the general need for finality
in judicial proceedings. If the reports had been
admitted it is clear that the
court would have had to have acceded to the Council’s request to re-open
the trial. At the re-opening
one would expect that the first contention would
have been that there were not in reality three new bursts, but only one, the
second
and third following shortly after the first because of steps taken to
isolate the first. Then there would no doubt have been evidence
as to bedding
conditions, and so forth, ad nauseam. The Council was entitled to have
an end to these already very protracted proceedings, and Mr Smit, for Mostert,
fairly conceded
that if the trial would have to be re-opened, his client also
would rather have a final decision based on the existing record.
[47] The
appeal is dismissed with costs, such costs to include those consequent upon the
employment of two counsel.
W P SCHUTZ
JUDGE OF APPEAL
CONCUR
SCOTT JA
MTHIYANE
AJA
SAFLII:
|
Terms of Use
|
Feedback
URL: http://www.saflii.org/za/cases/ZASCA/2000/35.html