222 MESSRS. C. T. HEYCOCK AND F. H. NEVILLE 



the door-panel moulding described later as characteristic of the E body. This alloy 

 polishes to a fine white surface. We must have passed the intersection D of the 

 branches CD and DE of the curve. The curve and the microscope here, as usual, 

 agree very well, the upper fugitive freezing point being due to the small amount of 

 the new primary crystals, and the lower very steady freezing point to the solidifica- 

 tion of the large mass of mother-substance. As by adding aluminium we now pass 

 up the curve towards E, the successive alloys show the primary crystals increasing at 

 the expense of the ground, the 29 '9 atom alloy being a good example (fig. 17), until 

 at 32 '5 atoms we have the new body in polygons separated from each other only by a 

 fine ribbon network of a different material (fig. 18); and at the point E, close to 

 33*3 atoms, we have a practically pure substance. 



The D and E bodies are white, both before and after etching, and each appears in 

 two forms, according to the orientation of the crystals to the plane of section ; one 

 form is smooth, like milky ice or ivory ; the other is uniformly pitted, or in some 

 cases ruled with lines. It would seem that the crystals are made up of minute rods 

 closely packed together,* and that section by a plane parallel to the rods gives a 

 smooth surface, perhaps a cleavage plane of the crystal. On the other hand, a section 

 making an angle with the rods breaks off each rod along another cleavage plane, 

 which will not generally be that of the section, in which case the surface will be 

 serrated or pitted. This way of accounting for the markings on a section of a mass 

 of crystals has no doubt occurred to many persons ;t but it was suggested to us by 

 Professor EWING'S slip-lines. He has shown that the minute elements of a crystal, 

 even of metal, are rigid, and that they can only slip along certain planes. In the 

 same way it is almost certain that they can only break along certain planes. 



The remarkable effects that one often sees on rotating a section under oblique 

 light are due to these serrations.J The fact that the pitted crystals give this rotation 

 effect, shows that the sides of the pits have a definite orientation. The variation in 

 the appearance of different crystals of the same substance, dependent on their orienta- 

 tion to the plane of polishing, is liable to mislead anyone who trusts to microscopic 

 study only. For example, the aUoy with 2 7 '2 atoms photographed with normal 

 light (fig. 15) appears to consist of at least two materials, while in reality all the 

 patches in it are of the same substance ; oblique light would be even more deceptive 

 as the slowly-cooled alloy with 26'1 atoms (fig. 12) proves. 



* An examination of the door-panel moulding of the alloy with 33 - 6 atoms almost forces one to the 

 conclusion that these minute rods or laminae came into existence during the crystallisation, and are not 

 a later product due to strain or other cause ; that they are, in Tact, the crystals of which the blob or 

 polyhedron is built up. This is the view insisted on by Mr. STEAD (loc. cit.). 



t Since writing the above we see that this is the view of the phenomenon given by Professor ARNOLD 

 ('Engineering,' February 7, 1896). 



J The nature and cause of this change of appearance when a section is rotated under oblique light is 

 very clearly explained by Mr. STEAD (loc. cit.). He attributes the first mention of it to Professor ARNOLD. 

 We shall call it the rotation effect, as no name appears to have been given to it hitherto. 



