I never saw any twist-drills in England except at Mr. Whitworth’s, and these I thought were the funniest things I ever did see. They were twisted by the blacksmith out of square bars and with a uniform quick twist, were left rough, and did not fill the hole, and the ends were flattened out in the form of the common drill to scrape, and not to cut. When I returned from England in 1868 twist-drills were coming into general use in this country. After 1876 the firm of Smith & Coventry introduced them in England. At that time almost everything in machine-shops was done in the old-fashioned way, and accuracy depended entirely on the skill of the workman. The tool work left much to be done by the fitter. Interchangeability was unknown, even in screw-threads. For example, when nuts were removed from a cylinder head, pains had always to be taken that each nut was replaced on its own bolt, as no two were exactly of a size. This condition developed a class of very skillful all-round workmen; but my earliest observation showed me that in manufacturing it was important that so far as possible the personal factor should be eliminated. I adopted the rule that in mechanical work there was only one way to insure that anything should always be done right, and that was to make it impossible that it should be done wrong. For example, in my governor gears their true running required that the bore should be absolutely correct, both in position and in direction. I had seen many gears bored. They were held in the jaws of a chuck and trued by marking their projecting side when running with a piece of chalk. It was evident that absolute truth could hardly ever be reached in this way, and the approximation to it depended wholly on the skill and pains of the workman. Besides, much time was lost in setting each wheel. These objections were much aggravated in the case of bevel-gears. I met these difficulties in this way. In standardizing my governors I found it necessary to make eight sizes, but managed to use only three different pairs of gears. I made a separate chuck for each of these six wheels, the faces of which were turned to fit the top and inner ends of the teeth, the same surfaces to which I had seen the chalk applied. When the castings were received from the foundry the first operation on them was to bed them to their chucks, which were covered with a thin coating of red lead for this purpose. The workman was careful to remove only projecting imperfections without touching the true surfaces of the teeth. After this the gears, being held firmly to their chucks by means of a yoke, were bored rapidly and always with absolute truth. Result: their running was practically noiseless.