Sir

I am very much in favour of Nature's use of Scientific Correspondence to bring to our attention potentially important research before full publication of results. But there is a need for scientific rigour in the presentation of the information to ensure that it is not misrepresented. The report by Losey et al. in Scientific Correspondence is a preliminary research finding offering insight into the topical and important issue of genetically modified (GM) crops (Nature 399, 214; 1999). It has alerted the UK Advisory Committee on Releases to the Environment (of which I am chairman) to a potential problem that will require very serious thought.

It has also been assumed to demonstrate that GM crops harm butterflies and has fuelled public anxiety about such crops. The report is highly indicative of harm caused by pollen from the maize plants used in this particular study, but the work is at this stage preliminary rather than definitive. It is clearly stated in the text, for example, that pollen for the non-Bt (Bacillus thuringiensis) maize control was from an unrelated, untransformed hybrid. However, there is no reported control to demonstrate that pollen from the transformed variety was not toxic in the absence of the functional Bt gene.

Losey et al. refer to a study (their ref. 3) stating that maize pollen is dispersed over at least 60 metres, and comment that a substantial portion of available milkweed (eaten by monarch butterflies) may be within the range of corn pollen distribution. But Losey et al. do not report a dilution study, so there is no evidence that dilute pollen would have an impact on larvae. Of course it is desirable to point out the potential harm that may arise from pollen dispersal, which in this case could be very important, but the data reported by Losey et al. do not directly pertain to this issue.

On 20 May I stated on the BBC Radio Today programme that the letter by Losey et al. was not peer reviewed and that the work might be flawed. I am now aware that it was peer reviewed (as are all contributions to Scientific Correspondence), and wish to apologize. My suggestion that the work might be flawed was not intended as a slight but was a reminder to the press that preliminary observations should not be overinterpreted. Regrettably, most reporting of the communication has almost entirely ignored the need for such caution.