Contents - Faperta

Transgenic Resistance to Insects: Gene Flow 423
Geographic Isolation
Half of the pollen produced by an individual plant falls within 3 m and the probability
of fertilization afterwards decreases slowly along a negative exponential of the distance
(Lavigne et al., 1998). The percentage of pollen dispersal from B. napus in the center of a
nontransgenic crop has been estimated to be 4.8% (Scheffl er, Parkinson, and Dale, 1993).
In rape, pollen trap examination showed a close link between weather and pollen
amounts transported from rape fi elds by wind. The degree of outcrossing depended not
only on the pollen concentration in the air, but also on other factors such as the weather
and the abundance of fl ower-visiting insects (Saure et al., 2003). The frequency was estimated
to be 1.5% at a distance of 1 m and 0.4% at 3 m. The frequency decreased sharply
to 0.02% at 12 m and was only 0.00033% at 47 m. No obvious directional effects were
detected that could be ascribed to wind or insect activity. Pollen-dispersal distributions
based on dispersal from whole plots instead of individual plants might underestimate
the proportion of pollen that dispersed over average or long distances. For insectpollinated
outcrossing crops such as radish, strategies other than distance must be
employed to ensure complete isolation (Klinger, Elam and Ellstrand, 1991). Gaussian
plume models, which take distance and wind direction into account, have indicated that
small conspecifi c populations of Lolium perenne L. might, in some conditions, be swamped
by immigrant pollen, even if they are not directly downwind of the source (Giddings,
2000). In sunfl ower, gene fl ow decreased with distance. However, gene fl ow occurred up
to distances of 1,000 m from the source population, indicating that physical distance
alone may not prevent gene fl ow between cultivated and wild populations of sunfl owers
(Arias and Rieseberg, 1994). Pollen dispersal from transgenic cotton is low, but increases
with an increase in the size of the source plot (Llewellyn and Fitt, 1996). A 20 m buffer
zone has been suggested to limit dispersal of transgenic pollen from small-scale fi eld
tests. The border rows fulfi lled the purpose of serving as a pollen sink to signifi cantly
reduce the amount of pollen dissemination from the test plot of cotton (Umbeck et al.,
1991). In China, gene fl ow between cultivars of G. hirsutum is up to 36 m. A buffer zone
of at least 72 m has been proposed as a safer distance to avoid gene fl ow. A minimum
isolation distance of 1,557 m has been recommended to prevent gene fl ow in alfalfa.
Complete containment of transgenes within alfalfa seed or hay production fi elds would
be highly unlikely using current production practices.
Use of Border Rows
Barren zones of 4 to 8 m may actually increase seed contamination over what would be
expected if the intervening ground were instead planted entirely with a trap crop.
When trap crops occupied a limited portion of the isolation zone separating transgenic
and nontransgenic varieties, the effectiveness of the trap crop depended on the width
of the isolation zone. Gene escape was reduced when the two varieties were separated
by 8 m, but increased the gene escape across a 4 m isolation zone (Morris, Kareiva, and
Raymer, 1994). For relatively short isolation distances, the most effective strategy for
reducing the escape of transgenic pollen is to devote the entire region between transgenic
and nontransgenic varieties to a trap crop. There is a signifi cant reduction in pollen dissemination
as distance from the test plot increased (Umbeck et al., 1991). Outcrossing
decreased from 5% to 1% by 7 m away from the test plot. A low level (1%) of pollen
dispersal was recorded to a distance of 25 m. The border rows reduced the amount of
pollen dissemination from the test plot. Economic profi t can be maximized by removing