October i8, 1894J 



NATURE 



601 



crystals. No hoar-frost formed on the upper surfaces of 

 the velvet or cardboird, and none existed on the grass. 



IV. Artificial Hoar-frost. — Kxperiments hid been 

 planned before Christmas 1892, for the artificial produc- 

 tion of hoar frost It was thought advisable, before com- 

 pleting our arrangements, to search for any possible 

 traces of ho ir-frost in the refrigerating chambers used 

 for the frozen-meat trade in Liverpool. This visit ren- 

 dered e.\periments unnecessary, as it vielded a rich har- 

 vest of simple and compound forms of hollow pyramidil 

 crystals. All the variations observed in natural ho.u- 

 frost were met with, and the details of the forms were 

 registered by microphotographs taken with magnesium 

 light The ice chambers were cooled down to - 13' C 



Very large and beautiful simple hoppers were 

 obtained from ships used in the frozen-meat trade. 

 During the four to six weeks of transit from the River 



crystal. At the angles of crystals there is, for a given 

 area, a larger supply of material for growth than in 

 the middle of a side. Beautiful skeleton crystals of 

 potassium chloride can be formed by rapidly cooling 

 concentrated warm solutions. First, a great number of 

 micro-crystals are formed, which float about in the brine. 

 Any of these may form a centre of attraction round 

 which crystalline matter will aggregate. A small cube 

 will form the centre, and from each solid angle a straight 

 axial row of small cubes will arise.' The intervening 

 parts will gradually get filled up if sufficient time is 

 allowed. 



This type of skeleton crystal is evidently due to 

 overgrorvth. 



Quite differently formed, though with the same result, 

 are the hopper crystals of sodium chloride. NaCl is 

 almost equally soluble in cold and hot water. Unaffected 



Fr.. I. — Ice Slalagmitcs in the l-.iva Cavern ol .Surlsheliir (!<:• landl. 



riate to Liverpool the hold is cooled down to about - 5^ C, 

 and the enclosed air is perfectly calm. 



On the occasion of a visit to Berlin, in June 1S93, the 

 large cooling cellars connected with the " Muenchener 

 Brauhaus " were examined (by kind permission of 

 Director Arndt), and small hopper crystals were dis- 

 covered on the cooling pipes. 



V. Comparison with other Skeleton Crystals. — The 

 simple hollow hexagonal pyramids of ice naturally suggest 

 comparison with the well-known cubic " hopper " crystals 

 of rock salt and skeleton crystals of other substances. 



In crystal-building there is always a marked tendency 

 towards ex cessive growth along thediagonal axes of the 



1 Our thanks are due to Mr. Ward for permission to work in the chambers 

 of the Sausincna Company, and to Mr. Lintott, through whose help we were 

 enabled to exhibit some of the ice crystals at the soir<^e of the Royal 

 Society. 



by cooling, it will therefore crystallise out of brine most 

 rapidly where the concentration becomes greatest through 

 evaporation, viz. at the surface. Suppose, then, a single 

 cube to be formed at the surface. Beginning to sink, 

 there will be deposits of fresh cubes on the four 

 upper edges of the cube in the form of a step. This goes 

 on until we have a floating hollow pyramid, apex down- 

 wards. At the corners of these hoppers additional cubes 

 are formed. (Compare the analogy with the hexagonal 

 form, Fig. 3, f.) These skeleton crystals are due Xo growth 

 at the upper edges of floating crystals. 



A third type of hollow skeleton crystals we have in 

 hoar-frost. 



When crystallisation of atmospheric vapour takes 

 place in absolute freedom, we find the crystals mainly 



1 SeeA. Knop. '• Molekularconstitution und Wachsthum der Krysulle," 

 p. 5J. (Leipzig: J867.) 



NJ. 1303, VOL. 50] 



