As the electrical industry grew, the need arose for some reliable form of meter to measure the amount of energy supplied and to form a basis of charge which would be acceptable to both supplier and consumer. Ferranti's first meter of 1882 was an electrolytic device, but it is his invention (Patent 5926 of 1883) of the mercury-motor meter the following year that is important, since this led to the first meters which were commercially successful and these were by far the most widely adopted in Great Britain for measuring direct current.
Ferranti himself gave an interesting account of his mercury-motor meter and its evolution in an address to the Royal Scottish Society of Arts at Edinburgh in 1895: 'The meter depends for its action upon the principle of a small motor driven by the current which is to be measured.... Its action depends upon the rotation of a radial current in a magnetic field. . . . The simplest possible meter ... consists of a coil wound of copper strip, into the last turn of which a metallic ring is soldered, this ring forming the sides of a small bath; the top and bottom are formed by pieces of ebonite or similar insulating material. This coil of copper ribbon forms the circular current, and in order to provide a conductor for the radial (or repelled) current the bath is filled with mercury, which makes contact at a metallic stud attached to the centre of the lower block, and also at the metallic ring. In this case we have to consider the radial current which flows in the mercury to be split up into many radial currents flowing from every part of the metallic ring to the centre of the bath. The mercury, when the current passes, is thus caused to rotate by the principle above described....
In the very earliest form of meter made, there was no magnetic circuit of iron provided to concentrate the lines of force generated by the circular coil, and cause them to pass through the mercury bath. The meter gave in consequence a very poor return in the way of power for the amount of electrical energy that it absorbed. Hence, when the current was small, the registration of the meter was both uncertain and unsatisfactory.... An iron pole was surrounded by the copper coil, and round about this another iron tube was arranged.... Now, when the current passes round the copper coil, it produces a powerful magnetic field in the pole piece, making the top side of the iron tube and bottom of the bath N and S pole respectively. Thus the lines of magnetism ... are more or less concentrated and caused to pass through the mercury bath.
This effected an improvement, but still was not satisfactory, as the air gap through which the lines had to pass was very great. At that time, also, the fans were not completely immersed in the mercury,... and it was found that in order to make the mercury start rotating, it was necessary to pass 20 amperes. Hence it was absolutely bad from the point of view of being a commercial recorder ... I was exceedingly disappointed. I could not understand why it did not go round when, as I thought, it should have done. I took the fan out and stirred up the mercury, and I found it ran round at first, and then stood still when it had about 10 amperes passing through it. I also noticed that although the surface of the mercury stood still, that it showed waves upon it. This showed me very soon, after making a few more experiments, that the non-rotation of the mercury was due to the surface tension of the mercury when exposed to the air.
I remember explaining this action to Lord Kelvin very shortly after these experiments were begun, . . . and he said, ... "you might as well have tried to sail a boat in a frozen loch, as get your meter to rotate with the fan dipping through the surface into the mercury below." It was, of course, a very good explanation, perfectly correct in every sense.' Ferranti then went on to relate that, after immersing the fan completely and making some further modifications, the starting current was reduced from 20 amperes to 0.3 ampere. This was really the beginning of the meter in practical form and it was followed by many other developments in meters for direct-current and alternating-current supplies.