A brief discussion is given of the need for pressures below 10−9torr and of the developments which began in 1950 and made possible the reliable measurement of pressures as low as 10−12torr. The greatest technical change in passing from high to ultra-high vacua is in materials since all components exposed to the vacuum have to be baked to at least 300 °c. All of the pumps, gauges, mass spectrometers, taps and other components described in this article must fulfil this requirement. Examples are taken from two small ultra-high vacuum systems in use at Southampton, one is a getter-ion pumped system in stainless steel, the other a mercury-diffusion pumped system in glass.