Carbon dioxide is often used as an indicator for the sensoric load from persons. The paper shows the result of measurement of carbon dioxide distribution in a test room with natural ventilation and high occupant density. The results show possible measuring positions for carbon dioxide in classrooms for controlling indoor air quality. The testroom represents one tier of a typically deep classroom with single - sided natural ventilation. The window was opened and closed automatically. Dummies emitting heat and carbon dioxide simulate the pupils. The measurements of carbon dioxide were carried out at various places and heights with a photoacoustic multi-gas monitor. With the window closed and no carbon dioxide emitted, the concentration of carbon dioxide equals in the room in a few minutes. With an open window and carbon dioxide emitted, the distribution of the gas varies. The highest concentration is below the ceiling, the lowest above the floor. When there is no thermal buoyancy at the dummies, more carbon dioxide emitted from the dummies drops down to the floor because of its higher density. Near the window as well as below the ceiling high fluctuations of carbon dioxide concentration occur. The measured values near the wall are more stable than the values in the middle of the room. The concentrations in the breathing zone show nearly the same values than in the middle of the room. The layered arrangement of carbon dioxide at the wall was not as clear as in the centre. The distribution of carbon dioxide is influenced both from the opening of the window and from the disposal of carbon dioxide. The concentration of carbon dioxide therefore varies within a wide range. Appropriate measuring points for classrooms could be at the wall at breathing height with an adequate distance from the windows. Higher measurement points at the wall, out of the students reach, could also be possible, but may show too high values. Measurement points near the floor and near the windows are inappropriate to estimate indoor air quality.