Monarchs and GM crops

Ian Harvey harveyi at liverpool.ac.uk
Fri May 21 11:02:39 EDT 1999


Dear all

In all the debate about the possible dangers of GM crops, it seems to me
that scientific evidence should be of the highest quality.  I'm afraid the
Nature paper which prompted the recent discussion of GM pollen and Monarchs
is very poor.  If it was an undergraduate project it would get a lower
second class mark.  I would point out three difficiulties with the study:

1. The 'dose' of pollen is unquantified. We are merely told that 'Pollen
density was set to visually match densities in milkweed leaves collected
from corn fields'.  Surely referees should have insisted on a quantitative
estimate.  It wouldn't be difficult to do.

2.  No proper control was conducted.  Pollen was used from the GM corn and
an '...unrelated, untransformed hybrid,..'.  Leaving aside the lack of
information as to the identity of the other strain, it is clear that this
also caused a reduction in feeding (but not more mortality).  There is no
evidence that it was the Bt transformation that caused mortality and
reduction in feeding: it could have been other charactersitics of the
transformed strain.  Either the untransformed strain should have been  used
as a control, or pollen from several other strains tested to see if the Bt
strain had a significantly greater effect.

3.  In total 75 larvae were used from a lab culture.  Given the importance
of this stud(it made front pages of several UK newspapers, and was reported
by all the main national titles), I'm convinced a more detailed study
should have been conducted.  Larger sample sizes, dose-response curves,
interactions with other factors (such as application of pesticides) all
need to be investigated.

To me, the Nature paper demonstrates two things:

a) Monarch larvae don't like corn pollen

b) Nature will publish insubstantial science to get in the press.

Finally, apologies to the authors of the work: it's the fault of the Nature
editorial process.

Cheers

Ian

PS I just quoting P-values isn't good enough: we should be told the
statistical tests they used.

******************************************************************************
Ian Harvey                                             Tel: +151 794 5028
Population Biology Research Group                      Fax: +151 794 5094
School of Biological Sciences, Nicholson Building
The University of Liverpool                            email:harveyi at liv.ac.uk
PO Box 147, Liverpool L69 3BX, UK
******************************************************************************



More information about the Leps-l mailing list