January i, 1920] 



NATURE 



449 



ments, under the latter cuboidal crystals, while the 

 lowest stratum consisted of the eutectic. It was evi- 

 dent that the compound richest in arsenic was the first 

 to freeze and floated upwards to the surface. As yet 

 the analyses have g'iven no decisive results as to the 

 composition of these shells. Further details have been 

 promised by Or. .Stead, and will be published in due 

 course. 



FiG. I I Nature-prill ). 



la the instance just quoted, four distinct stajjes of 

 crystal growth can be observed. When, however, the 

 proportion of antimony is between 20 and 25 per cent, 

 and that of arsenic about 5 per cent., the primary 

 crystals are distributed evenly through the whole alloy 

 and there is no stratification. 



That the primary crystals which form in such alloys 

 are splierical shell crystals was shown by chilling 



Fn;. 2 t( Photograph). 



them just l)elow tlie first thermal arrest. Kig. 2 

 depicts the structure after this operation. Sections 

 of the shells are visible which are smooth on the 

 ri)n(a\-.', but slightly broken on the conve.\, side, 

 dur: possibly to the stresses set up during quenching. 

 With somewhat slower cooling, as shown in 

 Fig. 3, both surfaces of the shells are seen to be 

 smooth. That these are composite and contain a 

 NO. 2618, VOL. 104] 



hard primary core was shown by grinding and polish- 

 ing experiments. 



In the latter part of his paper Dr. Stead quotes the 

 opinion of Mr. L. J. Spencer, to whom specimens of 

 the alloy containing spherical shells were submitted 

 for his opinion, and who furnished Dr. Stead with 

 important data regarding the curvature of crystals in 

 minerals. Mr. Spencer's complete notes on the subject 

 have been communicated to the Mineralogical Society. 

 In them he refers to various instances of apparent 

 curvature classified under these headings : — 



(i) Curved crystallites; (2) capillary habit; (3) ag- 

 gregations of crystals; (4) interfacial oscillations; 

 (5) vicinal faces; (6) bent crystals; (7) twisted crystals; 

 and (8) cylindrical and spherjcal crystals. 



It would appear that, according to his view, the last- 

 named constitutes the closest analogy to the alloy in 

 question. "The mineral kylindrite is a sulphur salt 

 of tin, lead, antimony, and iron. ... It has the ap- 

 pearance of consisting of tightly wound rolls of foil 

 with a smooth surface and a brilliant metallic lustre. 

 The ore consists of large numbers of these rolls, with- 

 a more or less radial grouping. . . . The rolls have a 

 diameter of a few millimetres up to one centimetre, 





■ • ■^•"^•w^^'i^'*™"'' 



., r>^^JvE?'^ 



Fig. 3 (Photograph). 



and reach a length of three to four centimetres. They 

 flake oft in concentric cylindrical shells with all the 

 appearance of a perfect cleavage, very similar to that 

 of the allied minerals, franckeite and teallite. These 

 cylindrically curved cleavage flakes are perfectly 

 bright and smooth and show no visible signs of being 

 built up of smaller elements. Spherical aggregates of 

 crystals' possessing a perfect cleavage are, however, 

 met with, but here a radial grouping is much more 

 coinmon than a concentric arrangement. Examples of 

 radiating spherical aggregates of lamellar crystals with 

 platy cleavages are pyrophyllite, zeophyllite, gyrolite, 

 faroeiite, tyrolite, etc." 



With reference to the cases of curvature in minera,! 

 crystal's thus referred to. Dr. Stead contends that none 

 approach in character or form the spherical shell 

 crystals obtained in his ternary alloys ; that radial 

 crystallisation round a nucleus is common in 

 minerals, but the spherical form finally produced is 

 an aggregation of manv crystals, and not a single 

 crystal ; and that kylindrite consists of a number of 

 cylindrical crystals which have formed round a central 



