Some methods of measuring the attenuation in waveguides are discussed, and the particular difficulties involved in the determination of low attenuations are outlined. The measurement of the Q-factor of a resonant line appears to be the most satisfactory method for short specimen lengths of waveguide, and the paper deals chiefly with the special equipment required for the measurement of Q-factors of the order of 106. The frequency stability required for the measurement of the very small frequency changes involved is very considerable, and the apparatus described ensures that frequency instability of the oscillator will not affect this measurement. Frequency differences are measured by a frequency discriminator using a waveguide delay line, the output of the discriminator being capable of accurate absolute calibration.The effects of cavity losses other than wall losses are theoretically investigated, since their magnitude influences the minimum length of cavity that can be used.