Abstract To study the influence of wind speed on ventilation rates, experimental trials were carried out in a plastic tunnel with continuous side ventilators. This type of greenhouse is the most representative of Mediterranean and arid regions. The injection peak technique using N2O as tracer gas was used to measure air exchange rates. Simultaneously air pressure difference between inside and outside at ground level was recorded together with outside wind velocity and direction and inside and outside air temperatures. Several combinations of side ventilator openings were studied under different wind velocities and directions. Results from pressure measurements show that the thermal buoyancy effect has no influence on the pressure field when the wind velocity exceeds 1·5 m s−1and that the turbulent wind pressure coefficient increases when adjacent obstacles perturb the wind. Results from gas measurements show that the air exchange rate is strongly dependent on wind velocity and total ventilator area, as expected, and that it can be expressed as a function of a global wind coefficient having properties similar to the turbulent wind coefficient. This indicates that the induced turbulent component of the air exchange rate can be significant. The reduced form of the ventilation flux, defined as the ventilation flux per unit ground area and average wind speed is used to compare the present results with those of other authors on different types of greenhouses and ventilation systems. It appears that roof openings are more effective than side openings and that the effectiveness of the ventilation vents depends on the type of opening. The effectiveness appears to be less when the vents are equipped with a side-mounted flap on their frame.