The aim of this paper is to analyse the effect of adjuvants and formulations on drift. The spray liquids consisted of four adjuvants (Actirob 0.4 %, Tensiofix 0.2%, Breakthru 0.2%, Silwet L-77 0.1%) with water and with two formulations of Phenmediphame (C16H16N2O4, 4.45%): an emulsion-forming (EC) and a suspension concentrate (SC). A standard flat fan nozzle at a pressure of 3 bar was used. The droplet size spectrum of each combination was determined using a Malvern granulometer. The droplet size was characterized by the volume median diameter (VMD) and the percentage of spray volume contained in droplets <100 microm (%<100). The relative drift potential was measured for each combination of formulation and adjuvant in a wind-tunnel. This latter has a working section 2.0 m wide by 2.0 m high by 6.0 m long. The air-stream is drawn by a 1.2 m diameter axial flow fan, powered by a 22 kW electric motor. Wind speed was 5 m/s. Its uniformity was controlled by a three-dimensional sonic anemometer able to move on a linear translation beam placed in the tunnel cross section. The wind-tunnel was operated under ambient conditions and three repetitions were performed randomized in order to eliminate variations in temperature and humidity for each combination. The ground spray deposits were measured on glass fibber collectors using a fluorescent tracer dye (sodium fluorescein), at a concentration of 2.5 mg/l. The statistical analysis of the droplet spectrum showed that the Phenmediphame SC formulation generated droplets of higher size than the EC. The mean VMD values were respectively equal to 228+/-11 microm and 185+/-11 microm for these formulations. For SC formulation, Break-thru decreased the VMD while Tensiofix increased the %<100. This confirmed that the degree to which an adjuvant influences spray characteristics is very variable. The drift profiles produced by the different combinations were similar, but the relative drift potential was significantly different comparing SC and EC formulations: it respectively reached 0.8+/-0,08% and 1.2+/-0,08%, whatever the adjuvant used in the liquid. Clearly, when using a flat fan nozzle to spray Phenmediphame, the droplet size and the drift potential are mainly governed by the kind of formulation, even if an interaction between the formulation and the adjuvant exists.