This report is mainly a discussion of dynamics of a cup anemometer. It is pointed out that the so-called overspeeding, caused by the atmospheric turbulence, is due to the fact that the cup rotor will speed-up more readily by an increase in the wind velocity than speed-down by a decrease of the same amount. The result is that, in principle, the measured wind speed with a given mean is larger than corresponding that obtained in a laminar (low-turbulence) wind tunnel with the same wind speed. This is a manifestation of the non-linearity of the cup anemometer response. In other words, the equation for the dynamics of a cup anemometer includes the wind fluctuations of second order. From a practical point of view the overspeeding is seldom important. Only when the wind speed is small, as in very convective conditions, it might be a problem. This is discussed in detail. The most important source of wind-speed bias is the non-ideal angular response. The cup-anemometer response to fluctuations in the (up-down) direction perpendicular to the rotor is in general important. To eliminate this bias it requires that the cup anemometer has a cosine response. Apart from this complication, the cup anemometer can be considered a spatial first-order filter, responding to fluctuations in the mean-wind direction, with a linear mean-wind calibration. The filter effect is described in terms of the distance constant, which is an instrument constant. The discussion is extended to the dynamics of instruments with non-linear calibration and instruments with second-order response.