When the stone-dressing machine was started a difficulty presented itself. The governor was in constant motion a short distance up and down, causing the engine to oscillate, running alternately too fast and too slow. There was nothing that should have caused this action, so far as I could observe. The load on the engine was constant. However the work done on the stone may have varied, the work of the engine was to lift the hammers, and these, being lifted successively, presented a uniform resistance. The oscillation was not very great, as nearly as I can remember about 12 per cent. of the speed; which would give to each hammer a variation of thirty-six blows per minute. This, however, produced a waving surface on the stone. The more rapid the blow, the stronger it was and the deeper the cut. These waves were slight, only about ¹⁄₅₀ of an inch variation in depth, but yet it was not possible for our rubbing-machine to grind them off without great loss of time. So we had to employ three or four stone-cutters to chisel off these ridges, which were about 4 inches apart. It was evident that this oscillation must be stopped. I tried to remedy it by changing the pressure of the steam, and then by changing the pulleys so as to run the engine faster, the speed of the governor, however, necessarily remaining the same. But these had no effect. Having exhausted my own stock of ignorance on the subject, I applied to professional experts for more, and I got it. Three persons, who I supposed ought to know, and who probably did know, all that was then known on the subject, gave me the same advice. It was that I should get a larger engine and a great deal larger fly-wheel. This advice did not seem to me reasonable. I knew that the engine was large enough, because while the governor was in the lowest position, in which it did not open the throttle entirely by any means, the machine ran too fast. They then told me I must have a heavier fly-wheel at any rate, and they explained to me that the fly-wheel performed two offices--one to carry the crank over its dead centers with an approximately uniform motion, and the other to give the governor time to act. I replied that the engine passed its dead centers with absolute uniformity then, as nearly as I could see, and as was shown by the surface of the stone, and consequently for that purpose the fly-wheel I had must be sufficient. The oscillations were regular, occupying about 30 revolutions of the machine, or 6 seconds of time, and had no connection with the dead centers, and I did not see why the governor should require any time to act. They told me that all governors required time to act, of course.