The same Bulkley pyrometer was used in all the furnaces to indicate the temperature of the escaping gases. On Tuesday morning, when my boiler was to be tried, I saw that before my arrival the pyrometer had been set in the brick chimney, where the readings could be conveniently taken by a person standing on the brick surface of the boiler chamber. Its readings averaged 260 degrees Fahrenheit. I did not believe this to be true. At about half-past two o’clock, when seven readings had been taken, one each half hour, having got ready some bricks and mortar and tools, I pulled the pyrometer out and filled up the hole. I then knocked a hole in the side of the brickwork at the bottom, in front of the flue, and set the pyrometer there. The reading rose to 405 degrees, which was the temperature at which the gases then entered the flue, and averaged about 385 degrees during the remainder of the sixteen readings. Root’s average was 416 degrees, and Phleger’s (also tubular) averaged 503. Obviously the readings taken before the pyrometer was moved should have been rejected; but the boys who did this kind of work added them all together, and our average temperature is printed 345.87 degrees, giving the boiler more credit than it was entitled to by about 40 degrees. I lost a little by this operation. While I was bricking up the hole the fireman came around and told me I was spoiling his fire. When I got the figures of water evaporated and coal burned, I found that in that half hour I had only 900 pounds (three barrels) credited to the boiler, instead of 1800 pounds (six barrels) during every other half hour, being a loss of about .023 in water weighed in the barrel, 38,400 pounds, instead of 39,300 pounds, while, curiously enough, the coal burned was rather increased.