Summary:

GM species are new, but historical parallels exist in the well-intentioned introduction of alien species to various ecosystems. In many such cases the results have been unpredictably disastrous.

Full comment:

I contribute as a professor in a physics department with interests in the modelling of complex systems.

Our ecosystem is extremely complex, and scientists who claim to know the full effects of introducing new organisms overstate their competence. Hence, each time a GMO is released, there will always be some risk of unforeseen ecological consequences. Such unpredictability is addressed, but not reduced, by theories of chaos, self-organised criticality, and other late 20th century advances in nonlinear science.

The risk is clear from previous instances where step-changes have been made in the genetic content of an ecosystem, with good intentions, using the best scientific advice of the day. In the past, this was done by the introduction of alien species. Such deliberate introductions (alongside many inadvertent ones) have sometimes been beneficial, but there have also been many catastrophes. See, e.g.:
www.gcrio.org/CONSEQUENCES/vol2no2/article2.html

In the GMO case the risks, though similar in character, are probably small. But few scientists (and fewer politicians) are trained to think objectively about small risks: unless we are very careful, even a modest economic benefit can appear to outweigh a small but real risk of catastrophic change. (Conversely, the small chance of a large gain can appear to outweigh a modest fixed cost, which is why Lotto is a success.)

A typical formulation in such cases is: 'there is no direct evidence of any such risk', or: 'any such risk is, at present, purely theoretical'. But theoretical risks of grave outcomes must be taken seriously: prior to Sept 11 2001, the risk of terrorists demolishing huge buildings in an American city was purely theoretical, and so, for a long time, was the risk that BSE could cause disease in humans.

In summary, past introductions of alien species to various ecosystems suggest that extreme precautions may be needed before the risks of further GM releases can be considered acceptable. (For example, released GM species should be incapable of sexual reproduction or hybridization of any kind, so that if an unforeseen ecological consequence occurs, the entire GM population can be extirpated without having permanently altered the gene pool.) Such extreme precautions may, of course, render most GMO technology economically non-competitive. In my view this is no reason to abandon those precautions.

Michael Cates
Professor of Natural Philosophy
School of Physics
University of Edinburgh