216 MESSES. C. T. HEYCOCK AND F. H. NEVILLE 



blobs being often oval or elongated into bars as with 28'6 (fig. 16) arid 29'9 atoms 

 (fig. 17). 



As we approach a minimum freezing point the rows of blobs become smaller in 

 area, and the mother-substance around them, when examined with a high power, is 

 sometimes seen to consist of a much smaller pattern of two substances, one being the 

 material of the blobs. The 38 '9 atom alloy (fig. 22) shows this well. Finally, at a 

 minimum freezing point, a eutectic angle, the large blobs disappear entirely, and the 

 whole alloy consists of the small pattern ; it is a eutectic alloy. The 40 atom alloy 

 (fig. 23) corresponding to the point F is an almost perfect example of this. If we 

 now, by the continued addition of the same metal, cause the freezing point to rise, 

 we again get large blobs surrounded by a minute pattern, but, while the fine pattern 

 is the same as before, the blobs are of a different material ; they consist of the second 

 substance of the network (fig. 24). Thus in crossing from one side of a eutectic 

 point to the other the two proximate constituents of the alloy exchange places. 



If we think only of the plane surface of such a section of alloy as that with 

 29 '9 atoms (fig. 17), the rows of blobs, each blob isolated from the next, yet 

 obviously connected with it by some law, are puzzling. But if we think of the solid 

 alloy as consisting of a mass of crystals with other crystals branching from them, the 

 whole system immersed in mother-substance which solidified after the formation of 

 the crystals, we see at once that a section of the mass would present the observed 

 appearance. To be more precise, we may picture the 29 "9 atom alloy during the first 

 stage of freezing as like a thicket of fir trees in which the branches are at right 

 angles to the stems and in which the stems are not all vertical. If a section were 

 made of this thicket by a plane inclined to the vertical we should get patterns very 

 like those of the photograph. If a stem lay in the plane of section we should 

 get lines at right angles to each other. If the stem were parallel to the plane, but 

 not in it, we should get parallel rows of dots. With the stem oblique to the section 

 we should get the elongated dots which are so numerous in the photographs. A 

 comparison of the X-ray photograph of the quickly cooled 96 '6 atom alloy with the 

 ordinary surface photograph of the same alloy illustrates the above. These con- 

 siderations, together with the straightness of the lines of dots, show that the large 

 pattern of blobs, and the polygons with which they are related, are the pure sub- 

 stance, which crystallised first and without constraint, because it was surrounded by 

 liquid, whereas the surrounding matter is a mother-substance that was liquid during 

 the first stage of crystallisation. It would be convenient to call the polygons and 

 blobs primary crystals, inasmuch as they formed first in order of time, while the 

 mother-substance may be said to contain secondary and tertiary crystals. It is 

 probable that the marked absence of crystal form observed in the blobs of primary 

 crystallisation is due to these blobs being what LEHMANN calls " crystal skeletons " ; 

 a snow crystal is the most familiar type of crystal skeleton. If we imagine the 

 interstices between the fern-like pattern of such a crystal to be filled up by sub- 



