meeting of the British Association at Ipswich in 1851. These were afterwards published in the Mining Journal. The consequence is that many of my Suction Ventilating Fans are now in successful action at home and abroad.

An Improved Method of Welding Iron.

One of the most important processes in connection with the production of the details of machinery, and other purposes in which malleable iron is employed, is that termed welding, namely, when more or less complex forms are, so to speak, "built up" by the union of suitable portions of malleable iron united and incorporated with each other in the process of welding. This consists in heating the parts which we desire to unite to a white heat in a smith's forge fire, or in an air furnace, by means of which that peculiar adhesive "wax-like" capability; of sticking together is induced, -- so that when the several parts are forcibly pressed into close contact by blows of a hammer, their union is rendered perfect.

But as the intense degree of heat which is requisite to induce this adhesive quality is accompanied by the production of a molten oxide of iron that clings tenaciously to the white-hot surfaces of the iron, the union will not be complete unless every particle of the adhesing molten scoriae is thoroughly discharged and driven out from between the surfaces we desire to unite by welding. If by any want of due care on the part of the smith, the surfaces be concave or have hollows in them, the scoriae will be sure to lurk in the recesses, and result in a defective welding of a most treacherous nature. Though the exterior may display no evidence of the existence of this fertile cause of failure, yet some undue or unexpected strain will rend and disclose the shut-up scoriae, and probably end in some fatal break-down. The annexed figures will perhaps serve to render my remarks on this truly important subject more clear to the reader.

• 
• Fig.1 represents an imperfectly prepared surface of two pieces of malleable iron about to be welded. The result of their concavity of form is that the scoriae are almost certain to be shut up in the hollow part, -- as the pieces will unite first at the edges and thus include the scoriae, which no amount of subsequent hammering will ever dislodge. They will remain lurking between, as seen in Fig.2. Happily, the means of obviating all such treacherous risks are as simple as they are thoroughly effective. All that has to be done to render their occurrence next to impossible is to give to the surfaces we desire to unite by welding a convex form as represented in Fig. 3; the result of which is that we thus provide an open door for the scoriae to escape from between the surfaces, -- as these unite first in the centre, as due to the convex form, and then the union proceeds outwards, until every particle of scoriae is expelled, and the union is perfectly completed under the blows of the hammer or other compressing agency. Fig. 4 represents the final and perfect completion of the welding, which is effected by this common-sense and simple means, -- that is, by giving the surfaces a convex form instead of a concave one.

  When I was called by the Lords of the Admiralty in 1846 to serve on a Committee, the object of which was to investigate the causes of failure in the wrought-iron smith work of the navy, many sad instances came before us of accidents which had been caused by defective welding, especially in the vitally important articles of Anchors and Chain Cables. In the case of the occasional failure of chain cables, the cause was generally assigned to defective material; but circumstances led me to the conclusion that it was a question of workmanship or maltreatment of what I knew to be of excellent material. I therefore instituted a series of experiments which yielded conclusive evidence upon the subject; and which proved that defective welding was the main and chief cause of failure. In order to prove this, several apparently excellent cables were, by the aid of "the proving machine," pulled to pieces, link by link, and a careful record was kept of the nature of the fracture. The result was, that out of every 100 links pulled asunder 80 cases clearly exhibited defective welding; while only 20 were broken through the clear sound metal. This yielded a very important lesson to those specially concerned.