We discuss calibration tests of our mobile Raman lidar intended for remote monitoring of atmospheric emissions from plant stacks. The calibration experiments have been conducted to validate its capability of remotely detecting, at the stack mouth, the most common gaseous pollutants like CO, NO, NO2, and SO2 using a specially designed gas chamber five meters long and 30 cm in diameter that can alternatively be filled with standard mixtures of the above contaminating species with pure nitrogen. The results obtained in these experiments well demonstrate the best features of the Raman-lidar technique in application to gas concentration measurements and this is illustrated with a perfectly linear concentration behavior of the Raman-lidar return from SO2 molecules in the concentration range of the latter from 0.05 to 5 % by volume. At the same time the experiments have clearly shown that such instrumental noises as long-lived broad-band luminescence of the chamber windows can impose a serious limitation on the possibility of certifying lidars as regards the minimum detectable concentration, since just this luminescence noise determines the signal-to-noise ratio at low concentration of a gas specie. In addition to this instrumental limitation there can occur certain troubles due to an unavoidable broad-band luminescence of a gas specie, like for example NO2, when using visible laser radiation to excite Raman effect in a gas mixture under study.