Large scale commercial tests of three infield refuge planting scenarios were conducted in 1998 on 2,500 acres of cotton near Vicksburg, Arizona to compare the development of pink bollworm infestation levels. Season long trapping monitored population dynamics. Single rows of conventional cotton were systematically planted within separate Bt-transgenic fields at the rates of one non-Bt row in four, six and eight rows of Bt cotton respectively. Each treatment was replicated six times.

Eighteen hundred (1,800) bolls collected on October 1 and 22 from non-Bt rows were incubated in boll boxes and third instar or larger larvae were counted. Results demonstrated that no differences exist between the medians of the three refuge treatments. A 36,000 boll sample collected from Bt replicates produced no larvae. Yield throughout the test area was well above the state average.

A second, non replicated test was conducted on Youngker and Youngker Farms in an area of much greater pink bollworm pressure in Buckeye, Arizona. This study suggested that crop losses incurred on an infield refuge of 25% non-Bt cotton produced less negative economic impact than chemical control costs on a nearby full field non-Bt external refuge crop.