
284 11. ENVIRONMENTAL IMPACT
with certainty an exact level of a chemical which
will cause one cancer in a million people exposed
to this chemical over a lifetime at low levels is not
being reasonable. Science is simply not that
precise by any means. For example, if
two
people
each smoke 40 cigarettes per day for 40 years,
about one million cigarettes are smoked. We
might expect one of these people to develop
cancer. In other words about one million ciga-
rettes might induce cancer in a person. Does that
mean a person who smokes a single cigarette in
his life has a one in a million chance of getting
cancer from it? We will never know until 25
million sets of identical twins are willing to subject
themselves to a life-long experiment where every
detail of their lives are controlled. However, it is
known that the risk of cancer in people who quit
smoking decreases significantly and after many
years approaches the risk level of people who have
never smoked, compared to their colleagues who
continue to smoke. This indicates that a certain
levels of toxins are required to induce cancer.
In the case of a compound like dioxin which
was found to be highly toxic in laboratory studies,
only later were studies done on humans who have
been exposed to it through industrial exposure
over many years, accidents, etc. Even these
studies are often done on small sample sizes by
people overly eager to advance their careers by
being able to report something "controversial".
By the time responsible, long term studies of
reasonably large populations are finished, the
damage has long been done, such as in the overly
high estimated toxicity of dioxin.
Third, some chemicals behave much differ-
ently in rats than in humans. The chemistry of
cancer is much more complex than the chemistry
of cyanide ion, which binds to hemoglobin and
prevents oxygen exchange by the blood where one
can expect the same result in a rat as in a human
as in any other animal with hemoglobin. But that
is certainly not true with long term carcinogens,
especially at very low rates of cancer mortality.
Fourth, the fact is that people in the medical
community tend to be aggressive; that is a factor
in the selection factor. Many studies get published
with statistical analysis, but statistically insignifi-
cant results. Let me explain. If one correlates 20
different would-be "dependent" variables to one
independent variable and finds there is a 95%
chance that one of
these
dependent variables corre-
lates to the independent variable, does this mean
anything? The answer is probably not since there
is a 5% chance that the correlation is just due to
chance. Since you started with 20 variables the
chances are that one of these 20 (which is 5%) is
bound to show a correlation just by chance at the
95%
confidence level. Stated another way, one
out of 20 researchers can expect to observe a
correlation at the 95% confidence level that is
merely due to chance for each correlation tested.
And these are the studies that use statistical analy-
sis!
Mark Twain is quoted as saying "there are
lies,
damn lies, and statistics!"
In the news media, which should not be
construed as scientifically (or otherwise) meaning-
ful,
one week coffee (caffeine) increases risk of
heart attach and cancer, and one week it does not
increase your risk. Along these lines, an article in
the New England Journal of Medicine (Fingerhut,
1991) indicates that earlier studies indicating the
toxicity of dioxin overstated the problem due to
small sample sizes and other reasons.
The EPA level of permissible dioxin emission
is 13 parts per quintillion in 1991; yet it is unde-
tectable below 10 parts per quadrillion (10,000
parts per quintillion). Do not shrug off these
incredibly small numbers without thinking about
them. One part in a quadrillion is only this much:
0.000000000000001! An olympic-sized pool
might have on the order of 250,000 gallons of
water or 1 billion grams. The level of arsenic in
drinking water is on the order of 0.01 parts per
million corresponding to 10 grams in the pool.
The measurable total amount of dioxin in the pool
corresponds to 0.00001 g; the proposed regulation
would be 0.000000013 g, 0.000013 mg, or 0.013
lig in the entire pool. (It would take more than 10
lifetimes for a person to drink the water in this
pool.) The area of low-level, long-term toxicity is
all very new ground. It is difficult to know what
rules,
regulations, and technologies will come out
of all of this. Certainly, however, it must be
handled in a more rational manner in the future.
11.2 WATER POLLUTION
It is useful to trace some of the important
U.S.
Federal legislation in the area of water. The