delivering 0.64 litre min À1 . The nozzle was positioned 350 mm above the laser crossover and each of three replicate measurements consisted of two sweeps through the spray cloud. In addition to the three oil emulsions, the droplet spectra from water, 1 g litre À1 aqueous Agral and aqueous emulsi®er at the three concentrations corresponding to the amounts present in the diluted oil ECs were measured. In addition, 1 g litre À1 Agral was added to 10 g litre À1 emulsions of MRO and MO to simulate the possible interaction between an emulsi®ed oil and a pesticide containing a wetting agent.
In-¯ight PDPA measurements of the spray cloud showed that all of the oil-in-water emulsions increased droplet volume median diameters (VMD) when compared with water and the aqueous surfactant solution (Table 1). For example, 10 g litre À1 MRO increased the VMD by 32% (247 to 326 mm) when compared with aqueous Agral. Emulsions of MO (10 g litre À1 ) produced greater volumes of driftable droplets than similar concentrations of the two vegetable oils (2.27% cf 1.57 and 1.58% for ARRO and MRO, respectively). The addition of 1 g litre À1 Agral to 10 g litre À1 emulsions of mineral oil and MRO increased the potential of both to drift but this effect was more obvious with the mineral oil than with the vegetable oils.
Good agreement was observed between in-¯ight droplet-size measurements and drift data for most of the formulations. Oil emulsions and the emulsi®er alone increased VMDs and decreased the proportion of driftable droplets with diameters less than 100 mm, compared with water and the surfactant solution. The surfactant generated the largest proportion of driftable droplets (5.75%) and 2.5 g litre À1 MRO the smallest (1.34%), in accordance with their wind-tunnel drift potentials.