Experiments on the mechanism and characteristics of a CO2‐CS2Geiger counter are reported, and circuits and methods of operation which satisfy the special requirements are described. Counter: The active part is 2⅞ in. in diameter and 16 in. long, filled to 1 atm, 95% CO2and 5% CS2. The cathode is copper and the anode is 0.005‐in.‐diam Chromel A. Characteristics: The electrons released by an ionizing particle become attached, probably to CS2. The self‐quenching action of the counter is excellent, no spurious counts are observed in the absence of electronic quenching. However, a dead time of several milliseconds must be imposed electronically because of the long interval during which the negative ions arrive at the anode. Method: An electronic quench is used which is triggered both by the CO2‐CS2counter and the anticoincidence ring. This serves to impose the required dead time and also to prevent the firing of the CO2‐CS2counter by mesons. Measurements: The plateau was measured to 1900 v above threshold and was found to be level to within 1% from 400 to 1600 v above threshold. The maximum drift time of the negative ions was found to be about 9 msec with a sharp cutoff. Tests with various combinations of gas indicated, but did not prove, that the charge carrier was CS2. The effects of common contaminants were determined. At voltages over 400 above threshold, 1% O2gave no detectable effect; 0.3% SO2gave a 2 to 3% reduction in counting rate. Extensive tests of the efficiency were made. Comparisons of the CO2‐CS2filling with an argon‐ethane filling gave identical results within the experimental error. No basis was found for supposing that there was any failure of the CO2‐CS2counter to register counts. Reliability: The counter described has been used for C14dating purposes over a long period of time. There has been no indication that there are variations in counting rate outside those expected on the basis of statistics. The counting rate for ``dead'' CO2in an iron shield 9 in. thick and after anticoincidence is 8 per min.