THE best method of testing telephone circuits and apparatus is a problem to which a great deal of attention has been devoted of recent years. Speech is carried over a telephone circuit by means of certain frequency components produced by the voice of the speaker at the sending end and received by the listener at the receiving end. If the components arriving at the listener's ear are exactly the same as when they left the speaker's mouth, the circuit would have a hundred per cent efficiency. There are two principal reasons, however, why the components arriving at the listener's ear are not the same as those leaving the speaker's mouth. In passing along the circuit the amplitude of the waves is attenuated; they thus become weaker. The amount of the attenuation also varies with the frequency, and hence the waves become distorted. Again, components which were not originally present are produced in the circuit. These are due either to noise or to the overloading of some part of the circuit. Their effect is to raise the threshold values at which the ear can hear different notes. In Electrical Communication for January, J. Collard gives a method by means of which the effect of given noises on the ‘articulation’ of a telephone circuit can be computed much more quickly than by the ordinary methods.