FOLLOWING the decision to develop a department of research and scientific study under A. P. M. Fleming, long-term investigations were held in abeyance by the war, and work was concentrated on manufacturing materials. A chemical section incorporating the early chemical laboratory was set up in June 1919 under R. W. Bailey, an early apprentice, and separate mechanical and metallurgical sections were established in the same year. A workshop for making experimental apparatus was started by N. Holt. New plant and instruments were hard to come by, and the arrival of a three-element oscillograph, still in use, was a major event.

In the insulation test the growth of work necessitated by 1920 a move to a larger area, and division into an electrical and magnetic section and a high voltage section; from 1919 this work had been in charge of B. A. G. Churcher, who later had as colleagues P. P. Starling and C. Dannatt and as apprentices D. B. Hoseason and J. D. Cockcroft, names that will come up again. Insulating materials today owe much to a close liaison between research department and the manufacturers. Before the first war, grey pressboard for use in transformers and other applications could only be obtained with difficulty from Germany, and the quality was uncertain. Methods of manufacture were investigated, and early in 1920 B. S. & W. Whiteley Ltd. agreed to produce it at their mill in Yorkshire. Thus a supply of high-grade grey pressboard was ensured. Between 1923 and 1925 much work was done on high grade porcelains and on micanite products, papers, and asbestos sleeving.

From the idea of material specification came that of process specifications. These were developed by the research department in 1921 as a step towards lower manufacturing costs, which would strengthen the competitive position of the Company. Process specifications define the technical operations to be followed in the manufacture or preparation of intermediate materials, ensuring that the materials are processed efficiently and are consistent in their properties.

Research into the magnetic behaviour of steels for electrical purposes was necessary to maintain the quality of incoming supplies of electrical steels and generator forgings and to supply the engineers with technical data, but new methods and apparatus for magnetic measurement had first to be devised. A permeameter of high accuracy was produced, and an iron loss tester (due to Churcher) having magnetizing windings that ensured a uniform flux distribution in the test specimen; both are now accepted as British standard testers and are used in many steelmakers' laboratories. Improvements in the wattmeter and bridge methods of measurement enabled iron losses to be investigated at frequencies up to 1 Mc/s and above.

Ways of physical testing that did not damage or destroy the specimen were much sought after. Methods of detecting cracks in iron and steel had been under investigation for two years when something unusual was noticed about a magnetized specimen that was being prepared for metallurgical examination: the very fine dust from the polishing operation always settled along a sharp curve.