The scientific underpinnings for the 1987 Montreal Protocol, the basis for current chlorofluorocarbon (CFC) phaseout policy, are quite shaky: the data are suspect; the sta- tistical analyses are faulty; and the theory has not yet been validated. The precipitous decision to stop CFC production by the end of 1995 was based mainly on press releases and panicky stories in the media about ozone holes, skyrocketing skin cancer rates, as well as blind sheep, the travails of plankton in the Antarctic and the demise of frogs in Oregon.

Linking the release of CFCs into the atmosphere with a putative increase of skin cancer involves a number of disputed steps:

CFCs with lifetimes of decades and longer become well-mixed in the atmosphere, percolate into the stratosphere, and there release chlorine.

Chlorine destroys ozone catalytically and thereby lowers the total amount of ozone in the stratosphere.

A reduced level of ozone results in an increased level of solar ultraviolet radiation reaching the surface of the Earth.

Exposure to more ultraviolet radiation leads to an increase in skin cancer rates and hundred of thousands of additional deaths.

It is well known that natural sources of tropospheric chlorine (such as volcanoes and ocean spray) are some four orders of magnitude larger than man-made sources. But it's what gets into the stratosphere that counts. Fruitless debate has focused on how much chlorine is rained out in the lower atmosphere; but is stratospheric chlorine increasing?

CFCs do indeed reach the stratosphere; the secular increase of fluorine--in the form of hydrogen fluoride (HF), as reported by Belgian researcher R Zander--is sufficient proof. But as late as 1987 (the year of the Montreal Protocol) Zander found no increase in hydrogen chloride (HCI), suggesting that stratospheric chlorine comes mostly from natural sources.

Earlier observations of HCI increases--between 1978 and 1982, by Mankin and Coffey at the US National Center for Atmospheric Research --had been used by some, even as late as 1993, to justify a CFC phaseout. This in spite of the fact that their data series was judged to be of poor quality and too short. Mankin and Coffey themselves ascribe their observed 1982 increase to the volcano El Chichon rather than to CFCs. The situation only changed in 1991, however, when NASA scientist Curtis Rinsland published data showing HCI increasing, albeit at half the rate of HE four years after Montreal.

The question of long-term global ozone depletion has been bedeviled by doubts about the quality of the data. For example, trends in sulphur dioxide pollution can be mistaken for ozone trends. More important, to establish the existence of a small, long-term trend it is necessary to eliminate the large natural variations, especially dependence on the 11-year sunspot cycle. This is an impossible task, given the shortness of the record and the virtual absence of data on long-term natural variations of the solar far-UV that produces ozone in the upper atmosphere. Not surprisingly, the claimed ozone trend fails a simple test; it depends on the choice of time interval.

Nor is the CFC-ozone theory itself in good shape. It was not able to predict the Antarctic ozone hole nor account for the reactions on volcanic aerosols in the lower stratosphere. Even in the upper stratosphere, where only gas-phase (homogeneous) reactions occur, the theory predicts much larger changes than are actually observed.

A major tear about a possible depletion of ozone--as expressed in the media--is that solar UV-B (280-320nm) radiation reaching the Earth's surface will increase, typically by 10% over the next 50 to 100 years, with dire consequences for plant and animal life and human health.

There has, of course, been a determined search for a secular increase in UV-B to match the presumed depletion of ozone, but no such trends have been observed. Recent claims of large trends between 1989 and 1993 over Toronto, Canada, proved to be spurious--an artefact of the statistical analysis. Moreover, average UV-B intensity increases naturally by about 5000% between pole and equator because of the change in solar zenith angle. Hence a 10% increase at mid-latitudes translates into moving just 100km to the south - hardly a source for concern.

Finally, much of the driving force behind the CFC phaseout policy has been the fear of an epidemic of skin cancer, particularly malignant melanoma. But unlike basal and squamous cell skin cancers, easily cured growths caused by long-term exposure to UV-B, melanoma rates do not show an increase towards lower latitudes. And indeed, recent laboratory experiments have established that melanoma is mainly caused by exposure to UV-A (320-400nm), which is not absorbed by ozone. Therefore, melanoma rates are not affected by changes in the ozone layer.

This important finding undercuts one of the main reasons for the Montreal Protocol and all subsequent efforts to phase out CFCs. haloes, methyl bromide, carbon tetrachloride, and other important halocarbons, including even CFC alternatives like HCFCs. With some difficulty, substitutes will become available. But with the health and environmental fears still unproven, the public may not take kindly to the huge cost--estimated at over $250bn worldwide--of replacing or retrofitting capital equipment that cannot accept the substitutes.