ON THE CHEMICAL NATURE OF ALLOYS. 185 



is almost universal. A few allovs, such as Levol's eutectic 

 copper-silver alloy, Cu 2 Ag 3 , and speculum bronze, SnCu 3 , 

 which Riche has shown to be free from a tendency to liquate, 

 show no structure under the microscope ; but all others can 

 be made by etching, or surface oxidation, to show two con- 

 stituents. This is very well seen in copper-silver alloys. 

 In pure copper the boundary lines between the crystals can 

 with difficulty be detected, but the addition of even a tenth 

 of a per cent, of silver produces a network of slender white 

 lines between the faces of the crystals of copper. These 

 lines grow in width with increasing percentage of silver, and 

 when the alloy attains the composition of the eutectic alloy, 

 about Cu 2 Ag 3 , they form the whole surface, the crystals of 

 copper having disappeared. With a larger percentage of 

 silver, dendritic crystals of this metal can be seen, so that in 

 all mixtures of these two metals we find an alloy of mean 

 composition filling up spaces between crystals of the metal 

 present in excess. 



Exactly similar phenomena are shown by bronzes, and 

 by many other alloys. It is hardly necessary to point out 

 that these facts agree very well with the views expressed in 

 the previous pages. Many of Behrens' results, however, 

 suggest the conclusion that the proximate constituents of a 

 solid alloy are frequently themselves alloys, and not the 

 pure metals. As an example of the complexity of the 

 phenomena observed by him, we give here his account of 

 the microscopic structure of a sample of gun metal. " Ten 

 per cent, bronze cast in bars. Well formed rectangular 

 rosettes composed of cubes of ioctyi in diameter. The 

 threads of light yellow alloy rich in tin, and measuring 

 20,u, often have an almost white central line ; moreover, 

 most of the pale red cubes are surrounded by a narrow 

 white edge, which stands out plainly from the pale yellow. 

 Spots of tin were not present." 



Steel, the most important of all alloys, will probably in 

 the end be found to have a structure similar to the above 

 described. Osmond finds nuclei of pure iron imbedded in 

 cells of a substance rich in carbon and silicon, and Sorby, 

 long ago, detected in many kinds of steel a microscopic 



