When ionization gauges are used as reference or secondary standards, their stabilities are the most important features. Generally, only hot cathode gauges are used as reference or secondary standards in calibration services. Some traditional reasons preventing the use of cold cathode gauges are the presence of discontinuities in the current versus pressure characteristic, the ignition delay at very low pressures, and the poor stability especially under contaminating vacuum environments. Though current inverted magnetron designs are believed to avoid some of these problems, at least under clean conditions, no "hard facts" have been published to support this claim. Here, we present the experimental results of calibrations of three hot cathode gauges and two inverted magnetrons against a primary high-vacuum standard over the range 10(-7) to 10(-3) Pa with gases of N-2, Ar, He, and H-2, respectively. During a continuous observation period of 72 h at a constant pressure of about 1 X 10(-4) Pa, hot cathode gauges showed better stabilities than inverted magnetrons in N-2, Ar, and He, but all gauges behaved similarly in H-2. Repeated calibrations over a period of about 6 months showed that all gauges had similar long-term stabilities in N-2, Ar, and He. For H-2, however, the stability of inverted magnetrons was better than that of hot cathode gauges. For different gases, the discontinuities of inverted magnetrons occurred at different pressures, making their calibrations more elaborate. (C) 2003 American Vacuum Society.