MINERALOGY AND CRYSTALLOGRAPHY 375 



taining 99-5 per cent, tin and 0*5 per cent, arsenic, crystals of 

 the intermetallic compound SngAsg separate in a pure condition 

 in a rhombohedral form with axial ratio a:c =1 : 1-2583. 

 From ternary melts, provided the ratio of antimony to arsenic 

 is greater than unity, the crystals are obtained in the form 

 of incomplete spherical shells, with diameter up to one centi- 

 metre, the best development of the curved habit being found 

 in melts whose composition lies within the limits 70-85 per 

 cent, tin, 25-18 per cent, antimony, and 4-5 per cent, arsenic. 

 The matrix in which the crystals occur consists of the antimony- 

 tin eutectic, but the arsenic-tin compound shows the same 

 habit when the above eutectic is replaced by the eutectics of 

 tin and lead or antimony and lead. The crystallographic details 

 in the paper are due to L. J. Spencer, who also summarises the 

 literature on curvature of crystals and its probable causes. 



The commonest type of curved crystal is probably the 

 microlites of hornblende, augite and so forth, which characterise 

 many vitreous rocks. In this instance the minerals have 

 developed in a solution of high viscosity, and the habit is due 

 to the prevention of diffusion, so that the crystals grow in 

 ever-changing directions, probably as the result of twinning. 

 Crystals with capillary habit — the so-called hair crystals — 

 possibly originate in a similar way. To this type belong the 

 hair crystals of silver and moss-gold {cf. A. Beutell, Cent. 

 Min., 14, 191 9), and possibly also the form of certain organic 

 compounds which show curvature when the crystals are 

 very thin, but not when they are thicker {cf. O. Lehmann, 

 Molekularphysik, Leipsig, 1, 374, 1888). In other cases the 

 curvature is due to the formation of an aggregate of minute 

 prismatic crystals which deviate slightly from parallel positions. 

 The formation of vicinal faces is probably responsible for the 

 habit of certain crystals of gypsum, diamond and so forth. 

 The curvature of quartz, beryl and tourmaline probably arises 

 through deformation along glide planes, while the twisted 

 crystals of quartz and dolomite are supposed to be heterogeneous 

 aggregates of small crystals. G. T. Prior {Min. Mag., 14, 26, 

 1904) has described a spherical habit in crystals of a complex 

 sulphide of lead, tin, antimony and iron. According to C. A. F. 

 Benedicks {Journ. Inst. Metals, 22, 145, 19 19) curvature is fairly 

 common in metals. For example, antimony often occurs in 

 curved form, while graphite in grey pig-iron develops in spherical 

 shells if the carbon crystallises out at a late stage in the cooling 

 when the metal is in the plastic condition, the curvature being 

 even more pronounced in the case of temper carbon which 

 forms when the iron is in a solid condition {cf. K. Honda and 

 S. Saito, Sci. Rep. Tohoku Univ., 9, 311, 1920). According to 

 Benedicks, where a solid solution is crystallising, not only do 



