Homage to Gaia (38 page)

My Green friends, many of whom believe in homeopathy and other kinds of alternative medicine, may be interested in my views on this kind of dilution. Homeopathic practitioners regularly serially dilute their drugs to levels trillions of times lower than parts per trillion. My experiences showed how difficult it is to dilute simple inert chemicals like CFCs. They made me deeply sceptical about the extreme dilutions of homeopathy. On the other hand, there can be few things as harmless as a drug applied homoeopathically, and there would be no side effects.

The other way to calibrate the
Shackleton
chromatograph was to make an absolute detector. This is an old trick of instrument scientists. It is simple to do, but it requires a theory of how the detector works. I assumed that the reaction between free electrons that were floating inside the detector and chlorofluorocarbon molecules led to their mutual removal from the detector. Each electron lost this way was then equivalent to one CFC molecule. I knew the average numbers of electrons in the detector quite accurately from the current flowing in it. A typical detector current of a hundredth of a microampere is exactly equivalent to 62.415 billion electrons per second. So if CFCs at a flow of ten billion molecules per second were passed into the detector, ideally they should remove ten billion electrons per second and hence decrease the current flow by 1.602 nanoamperes. From the simple arithmetic it was not difficult to calculate, from the area of a chromatograph peak, how many electrons were removed and hence how many CFC molecules had reacted. Reality is more
complicated
. Some of the CFC molecules could escape reaction and fail to be counted. To answer this, I joined two detectors in series and detected the amount of CFC that escaped through the first detector. It turned out that forty per cent of the CFC escaped the detector I was using. Therefore, I had to correct my first estimate by this amount. There were other more recondite doubts, mainly about other
reactions
of the CFC molecules in the detector. Nevertheless, I chose to use this absolute procedure to quantify my measurements made on the
Shackleton
expedition. This was both foolhardy and brave. First measurements have no predecessors to compare with. If I were badly wrong about absolute detection, I could not have chosen a better way to advertise my error than the first paper on the global distribution of chlorofluorocarbons that was published in
Nature
in 1973.

The first time that this procedure was called to account was at a meeting of the United States National Academy of Sciences’ panel on
ozone depletion. It was at the ski resort of Snowmass in Colorado, which is a beautiful place, at about 8000 feet up in the Rocky
Mountains
. One of the panel scientists asked, ‘How accurate are your
chlorofluorocarbon
measurements?’ ‘About twenty per cent,’ I replied. Twenty per cent accuracy implies that a quoted value of 100 parts per trillion represents a true value in the range 80–120. The errors start when the air sample is taken. How close in volume is the air sample to the 5 cc marked on the syringe used to collect it? They end with the estimate of the peak area of the chromatogram. On the
Shackleton,
I did it using a pencil and ruler. These errors added up, I calculated, to make an uncertainty of about twenty per cent. The true error may have been less but I had learnt in a lifetime of measurements that it is best use the worst estimate when trying to guess errors. Almost immediately, an American analyst present at the meeting jumped up and said, ‘Oh, I can measure the CFCs to a one per cent accuracy.’ I was impressed; I knew that my home-made apparatus and dubious syringe method of sampling were less professional than I would have wished. I had not realized how far ahead the high technology of the United States had gone. The panel was equally impressed and in its published report categorized my measurements on the
Shackleton
as inaccurate. It took me five years to discover that the claim of one-per cent accuracy from the young man who jumped up at Snowmass was false. The claimant, I discovered, was hazy in his mind over the difference between accuracy and precision. He meant one-per cent precision, not one-per cent accuracy. The difference is this: a badly inaccurate but precise weighing machine will record your weight as 90 lbs, never varying from 89 to 91, when in fact your weight is 150 lbs. An accurate but imprecise weighing machine will give weights between 130 and 170 lbs, and if you weigh yourself often enough will provide an average close to your true weight.

A professional body, the US Bureau of Standards, also grew
suspicious
of the analyst’s ability to measure CFCs with such astonishing accuracy. Ernest Hughes, William Dorko, and John Taylor of the Bureau designed an experiment to find out the truth about these claims. They filled one set of small gas cylinders with clean air and another set of cylinders from a batch of the same clean air with some pure nitrogen added to it. They sent one each of these cylinders to the principal analysts measuring CFCs in the atmosphere, and asked them to report their findings. When they returned their measurements, the National Bureau of Standards plotted the two
measurements from each analyst on what statisticians call a Youdon plot. This is a graph where the value of one measurement is marked by its position on the horizontal axis and the other measurement by its position on the vertical axis. If all the measurements were accurate to one per cent then all the reported values would have centred within a small circle—like the throws of a champion darts player, all in the bull’s eye. In fact, the FC11 results were scattered over a range going from less than half to more than twice the true value and the FC12 results were worse even than this. My twenty per cent did not look so bad compared with this. Their report, which revealed inaccuracy throughout the community, brought me sharply back to my student encounter with Professor Todd, and the time when he could not believe the accuracy of my student exercise, not knowing about my professional training. Scientists working at universities, and using unfamiliar techniques, often make inaccurate measurements because they rarely have the time to become proficient.

The National Bureau of Standards report in 1978 renewed my trust in absolute detection. I decided to validate it as a method. This I did by building a fifty-cubic-metre chamber, hermetically sealed, within an old barn deep in the Devon countryside. The idea I had was to prepare accurately small volumes of chlorofluorocarbon gas by a vacuum line procedure. Graham Milne, an ICI scientist, gave me the apparatus needed for this and generously gave his time teaching me how to use it. First, I prepared an accurately known volume of chlorofluorocarbon gas in a sealed glass ampoule and then took the ampoule to the barn and placed it before a powerful fan. I then left the chamber and set my instruments running, taking samples of the background air of the chamber, and then I broke the ampoule by an electrically driven crusher. The chamber air was continuously refreshed by air from outside at a rate of two air changes an hour. At intervals, samples of the chamber air were taken and measured. I checked the accuracy of this dilution method by also releasing
hydrogen
gas into the chamber, and following its dilution using a
well-calibrated
thermal-conductivity detector. Whilst I was doing these experiments, a family of barn owls took residence in the space above the chamber. The barn programme kept me busy for nearly three years, and it was some of the most exacting and least rewarding financially of my life’s work as a scientist. It was worthwhile because it did make honest the atmospheric CFC measurements, and it did confirm that the ECD could be an absolute detector for the
chlorofluorocarbons
.
Most of all, for me, it justified my estimate of the accuracy of the
Shackleton
data. The paper describing these
experiments
in the barn appeared in
Geophysical
Reviews
in 1984. For reasons I never understood, the ALE management team insisted that it appear as a joint paper with Rai Rasmussen, a member of the ALE team, as the lead author. Nevertheless, I did the barn
experiments
unaided by anyone.

During the time of the barn experiments my friend, Brian Foulger, and I made a series of measurements of the atmospheric abundance of halons, the bromine-rich compounds used as fire extinguishers. We made them here at Coombe Mill and in the southern hemisphere at Cape Town and in New Zealand. We found about 1.5 parts per trillion in the northern hemisphere and 0.6 in the southern
hemisphere.
The chemical company, ICI, funded the work. Unusually, but as was their right, they persistently refused us permission to publish these findings. This was the only time in my life as an independent scientist that a company or a government department blocked the publication of important scientific information.

The Ozone War was curiously involved with low-budget science done by British scientists in or near Antarctica. The voyage of the
Shackleton
to Antarctica in 1971 first drew attention to the global distribution of CFCs. The discovery by Joe Farman and Brian
Gardiner
of the thinning of ozone above the Antarctic landmass made us realize how serious it was. These were quiet, inexpensive researches inspired by a sense of wonder, and we were all inspired by the theory of Molina and Rowland in California, which was a modestly funded research. We need big science to complete our understanding of the intricate chemistry of the stratosphere, but we must never forget that pioneering small-scale research is just as necessary. I started the CFC ozone affair as small science, but by 1982 it seemed that I was on an accelerating bandwagon, and now was the time to jump off. At the end of that year, I fell ill and was unable to travel. I was glad that illness gave me the chance to escape what was now big science. I have never returned to it.

On the voyage of the
Shackleton
I had as a companion in the next laboratory on the ship a striking and friendly young German student, Hans Greese. His task on the voyage was the difficult one of
measuring atmospheric carbon monoxide. Like me, he wanted to know how it varied between the northern and southern hemispheres. My equipment was simple, home-made and occupied no more than four square feet of bench space. His was intricate and filled the whole of the front laboratory of the ship. I was fortunate to have the electron capture detector, which is specifically sensitive to the
chlorofluorocarbons
that I sought. Hans had to do it the hard way by extracting the tiny proportion of carbon monoxide, less than a part per million, from the air by standard chemical methods, and then measuring it by chemical quantitative analysis. Just as I was impressed with his
professionalism
and craftsmanship, he was impressed with the simplicity and sensitivity of my chlorofluorocarbon measurements. In some ways, it typified the different approaches of our two nations: the
German
as painstaking and professional and the English as opportunistic but effective amateurs. He must have talked with his colleagues and his supervisor, Dr Wolfgang Seiler, when he returned to Germany. The distinguished scientist, Christian Jünge, was director of the Max Planck Institute of Atmospheric Science at Mainz in Germany. He wrote to me inviting me to Mainz to tell them of the discoveries made during my voyage on the
Shackleton.

I travelled to Mainz sometime in the autumn of 1972. Dr JH Hahn met me at Frankfurt Airport and drove me by car to Mainz. They let me stay in the Institute guesthouse and Dr Hahn arranged to meet me there at 9 o’clock the next morning. After a good night’s sleep, I awoke hungry and ready for breakfast, and then a good day of talks about the atmosphere. I wandered around the guesthouse but seemed to be the only one there. There was no dining room and I could not smell coffee brewing or breakfast cooking. Hahn appeared sharp at 9 and I asked, ‘Are we going to have breakfast?’ ‘Breakfast?,’ he said with a grin, ‘you must have a German breakfast: a cup of coffee at the lab and a cigarette.’ This was a culture shock for me. Breakfast is the most important meal of the day for the English. I don’t go in for the whole thing: grapefruit and porridge, followed by egg, bacon, fried potatoes, tomatoes, and finished off by rounds of buttered toast and marmalade, all washed down with a large quantity of good strong tea. But to start the day on an empty stomach and no cup of tea was too much. I asked if there was a shop that sold something light. Jürgen Hahn kindly, sensing my lack of food, took me to a café where there were cakes and beer available. I settled for the cake and then returned with him to the Institute, where there was coffee brewing all of the time.

I was much impressed with the atmosphere and quality of the scientists there. Christian Jünge reminded me of Sir Charles
Harington
. He was a quiet, strong, and authoritative man with a
towering
intellect. We had a happy morning discussing the significance of the
Shackleton
analyses. Jünge had proposed in an earlier paper a way to estimate the atmospheric lifetime of a gas. He guessed that long-lived gases such as oxygen and nitrogen would not vary in abundance by a detectable amount, whereas short-lived gases like methane or carbon monoxide, with residence times of a few years or months, would fluctuate considerably in abundance. This approach suggested that the chlorofluorocarbons, in the southern hemisphere at least, had a long lifetime, which later we found to be over a hundred years.

Almost all of the Germans I met spoke good English and I felt ashamed at my lack of language. It was a pleasant stay and towards the end, Christian Jünge invited me to make another sea voyage, this time on the German research ship,
Meteor.
It was due to sail from
Hamburg
to Santo Domingo in the Caribbean in late 1973. What a wonderful ending for my visit: the prospect of another sea voyage filled me with joy and I now had something to look forward to and prepare for during the next year.