The use of a continuous air monitor (CAM) is mainly motivated by the need to be alerted quickly and in the most accurate way possible with an acceptable false alarm rate when a significant activity concentration value is exceeded, in order to take appropriate measures to reduce exposure of those involved. The performance of this CAM does not only depend on the metrological aspect characterized by the decision threshold, the limit of detection and the measurement uncertainties but also on its dynamic capacity characterized by its response time as well as on the minimum detectable activity concentration corresponding to an acceptable false alarm rate. The ideal performance is to have a minimum detectable activity concentration as low as possible associated with a very short response time, but unfortunately these two criteria are in opposition. It is therefore important that the CAM and the choice of the adjustment parameters and the alarm levels be in line with the radiation protection objectives. The knowledge of a few factors is needed to interpret the response of a CAM and to select the appropriate CAM type and its operating parameters. Among those factors, it is important to know the half-lives of the radionuclides involved, in order to select the appropriate detection system and its associated model of evaluation. CAM using filter media accumulation sampling techniques are usually of two types: fixed filter; moving filter. This document first describes the theory of operation of each CAM type i.e.: the different models of evaluation considering short or long radionuclides half-lives values, the dynamic behaviour and the determination of the response time. In most case, CAM is used when radionuclides with important radiotoxicities are involved (small value of ALI). Those radionuclides have usually long half-life values. Then the determination of the characteristic limits (decision threshold, detection limit, limits of the coverage interval) of a CAM is described by the use of long half-life models of evaluation. Finally, a possible way to determine the minimum detectable activity concentration and the alarms setup is pointed out. The annexes of this document show actual examples of CAM data which illustrate how to quantify the CAM performance by determining the response time, the characteristics limits, the minimum detectable activity concentration and the alarms setup.