_Internal corrosion_ sometimes takes place from bad feed water, and its effects are different in extent in the different parts of the same boiler. It very seldom thins the plate over a large surface regularly, but attacks the iron in spots, pitting it in a number of holes. These are sometimes large, as if gradually increasing from a centre of action; and sometimes small, but so close together as to leave very little more space whole than that which is attacked. A very curious example of the latter was exhibited to the meeting, and shown in Figs. 56 and 57, cut from the lower part of the shell of a large tubular boiler externally fired. The corrosion was greatest along that part of the shell most exposed to heat, and was so extensive that two boilers exploded simultaneously. The boilers had been at work sixteen years, but the corrosion commenced about eight years before the explosion, when the feed water was rendered corrosive by being obtained from some iron mines. This explosion occurred at Aberaman in 1864. The corrosion had been seen going on for years, and was not considered sufficient to cause danger; but the depth to which it extended through the thickness of the metal is seen in the half size section, Fig. 57. Another sample equally curious was exhibited to the meeting, and shown in Figs. 58 and 59, taken from the sweep plate over the fire in a plain cylindrical boiler which had worked about ten years. The feed water was occasionally bad, and attacked the iron over the area DDD, where unprotected by scale. The protection afforded by scale against occasional corrosive feed water is worthy of notice. In the two specimens exhibited it is seen that the protection has been perfect where the scale has not been chipped off; and the edge of the sound part projects over the hollow, as seen in the half size sections, Figs. 57 and 59, the corrosive water having eaten away a larger area beneath than that through which it first entered the surface of the iron.