544 



NA TURE 



[September 28, 1905 



(13) Sea ice and artificial ice break up in sunlight into 

 little hexagonal prisms of clear ice. These suffer mutual 

 displacement the less easily, the thinner are the fine foam- 

 walls (which have now melted again, and which, when 

 the freezing took place, were formed out of oily salt 

 solution, normal to the surface), and the less salt the 

 water contained before freezing. 



The purer the water was, the larger are these hexagonal 

 prisms or foam-cells. 



(14) The capillary fissures in transparent glacier ice are 

 these fine foam-w'alls of oily salt solution. 



{15) When water containing little salt freezes in deep 

 metal troughs surrounded with strongly cooled brine, the 

 oily salt solution separates in thin layers normal to the 

 surface, and forms bubbles, foam-cells clinging to one 

 another, or — when the oilv liquid at low temperatures is 

 very viscous — folds or hollow pipes, w'hich are filled with 

 pure or nearly pure ice, or with air if such were present 

 in the water. The artificial ice is seen to be traversed by 

 many horizontal tubes, normal to the surface, which are 

 specially numerous in the diagonal and median planes of 

 the ice block, where the mother liquor had accumulated. 

 The less salt is contained in the ice, the more transparent 

 are these diagonal and median planes of the artificial ice 

 block. 



Illumination with sunlight or daylight causes the appear- 

 ance of fresh tubes. The ice becomes more cloudy, and 

 subsequently more transparent again. 



(16) When water containing air freezes in deep metal 

 troughs, the upper part of the ice block shows horizontal 

 layers consisting alternately of transparent pure ice and of 

 opaque salt-containing ice with numerous air bubbles. The 

 more salt the water contains, the more numerous and the 

 closer are the opaque layers. In sunlight these opaque 

 layers melt more easily than the transparent ones, and 

 furrows are formed on the surface of the opaque ice. 



(17) If the ice is allowed to thaw again in a warm room, 

 or is exposed to radiation (daylight), the parts rich in salt 

 melt sooner than those which contain little salt. The 

 tubes of oily salt solution bulge and coil up, and then 

 break up with contraction of volume into spherical bubbles, 

 which may be vacuous or filled with air. The foam-cells 

 exhibit shapes like those of colloids and jellies as they 

 swell or shrink, or those tree-like and branched formations 

 which I have described in the case of the " liquid pre- 

 cipitates " of metallic silicates and cyanides. If the 

 capillary fissures in this opaque ice are filled with very 

 viscous salt solution, or if the oily salt solution forms no 

 continuous foam-cells, it cannot run away. The ice re- 

 mains white, as glacier ice actually does. 



(18) When an ice block thaws under the long-continued 

 action of daylight, there appear, in the diagonal and 

 median planes of the block, bright bands and cloudv bands, 

 which change their shape and position as the duration and 

 intensity of the radiation alter. This is due to the form- 

 ation of new foam-walls of oily salt solution and the dis- 

 appearance of old ones. The 'angles between the foam- 

 walls are also seen to change, which means that the surface 

 tension of these walls is changing. Now as the amount 

 of salt in the diagonal planes increases, and the absorbed 

 radiation diminishes, towards the interior of the ice, and 

 as further the surface tension and viscosity alter with 

 changing concentration and temperature, it follows that 

 the shapes assumed by the oily layers in the interior of the 

 ice under the influence of the surface tension also undergo 

 change. 



(19) After thirty to thirty-six hours, the block of artificial 

 ice had melted in the warm room to half its original 

 height (i metre), and at the foot and warmer places had 

 given w-ay in a pasty mass. In the upper portion, foam- 

 walls had formed in the pure ice, inclineti 120° to one 

 another. In these, as in the median layer that had thawed 

 away, melting salt solution ran down for hours. At the 

 warmer places, and at the thin uppermost crust, glacier 

 grains were formed. These w^ere foam-cells, 5 mm. to 

 10 mm. wide, filled with doublv refracting ice, and 

 separated from one another bv singly refracting foam-walls 

 of transparent salt solution. At the' junctions of the foam- 

 walls there often lay tetrahedra, bounded by spherical 

 surfaces and filled with transparent liquid. 



(20) In the diagonal and median planes of a block of 

 KO. 1874, \01. 72] 



artificial ice (i metre high) containing a certain very small 

 amount of salt, and exposed to a certain intensity of 

 radiation, there can be formed horizontal closed tubes of 

 pure or nearly pure ice, having rounded heads and sides 

 bulging at places, and filled with liquid salt solution. 

 They slowlv swell, slowly break up into separate bubbles, 

 and' then slowly pass away. They are first formed low 

 down, at places of high pressure, and afterwards higher 

 up, at places of low pressure. 



(21) When distilled water, free from air, was frozen in 

 iron troughs, it w-as found at a certain temperature or 

 with a certain concentration of the salt solution and the 

 oily foam-wall that the walls and contents of the closed 

 tubes in the lower part of the median plane were for 

 some time coloured yellow. Subsequently this colour dis- 

 appeared. It was not present when the water was frozen 

 in brass troughs. I believe it was due to ferric oxide, 

 which was differently soluble in the walls and in the liquid 

 inside the foam-cells, and at a higher temperature became 

 insoluble and sank to the bottom. 



(22) The phenomena of melting ice depend both on the 

 velocity of freezing and the velocity of thawing. The 

 more rapidly the water freezes, the more numerous are 

 the foam-walls, and the smaller the foam-cells. 



(23) Very dilute solutions of different salts, when slowly 

 frozen under similar conditions, give oily layers of varying 

 viscosity and surface tension, or spheres, bubbles, tubes, 

 and foam-walls of varying form. I have shown this with 

 freshly boiled water containing 0000003 per cent, ot 

 NaCI,' or equivalent quantities of KCI, KXOj, Na,SO,, 

 CaCI,, MgCU, Al„(SO,),. The water was "frozen in'pri- 

 matic troughs of brass or tin. 



(24) During the freezing of water containing 00015 per 

 cent, of Na.SO,, and also containing air, the air separated 

 at the same time as the mother liquor. The bounding 

 surface between air and almost solidified, very viscous 

 liquid, tends to become as small as possible, and rolls up 

 together to form hollow cylinders, the radii of which are 

 the smaller the more quickly the ice has frozen. The 

 water freezes the more slowly the further it is from the 

 strongly cooled (below 0°) side of the trough. The thin 

 layers forming the walls of the tubes are normal to the 

 solid surface of the side of the trough, or of the trans- 

 parent mantle of ice which encloses the mother liquor. 

 They frequently form cylindrical or conical tubes, 6 mm. 

 to 12 mm. long, with a whitish skin, and filled with air. 

 Their axes are normal to the surface, and their pointed 

 ends are directed towards the outer side of the ice mantle. 

 At the base of the tubes, which may be 0-5 mm. to 2 mm. 

 wide, there hangs a whitish hollow sphere inside the mother 

 liquor. 



(25) On slowly freezing water containing from 0-00014 

 per cent, to 0.0014 P^"" cent, of Na.SO,, or o 003 per cent, 

 of NaCl, it happens at times that the mother liquor, which 

 is surrounded by a transparent mantle of ice, contains 

 numerous flat crystalline plates of pure ice. These, by their 

 shape, position, and inclination to one another, clearly 

 show that they have been formed from thin oily foam- 

 walls of pure water, which, as the cooling proceeded, have 

 separated from the watery salt solution, and then solidified. 



(26) When a test tube, containing boiling distilled water, 

 is plunged into liquid air, the water freezes very quickly 

 to a milky-white mass of ice, with fissures normal to the 

 surface of the glass. If the test-tube with the white ice — 

 the whole being now cooled down to — iqo° — is plunged 

 into distilled water, it becomes coated on the outside with 

 a thin crust of ice, which can be detached with a knife, 

 and examined in a watch-glass under the polarising micro- 

 scope. It consists of small glacier grains or foam-cells 

 (01 mm. to 02 mm. in diameter) the flat walls of which 

 are normal to the cylindrical surface, and are inclined to 

 one another at angles of 120°, 110°, and so on. The interior 

 of each foam-cell contains a crystal of ice, which in the 

 different cells is differently orientated. When the ice in 

 the test-tube is crushed with a steel point, it exhibits a 

 fibrous fracture, with fine fibres normal to the cylindrical 

 surface. Occasionally in the cross-section are seen con- 

 centric cylinders composed alternately of transparent and 

 of white ice. The latent heat of the slowly freezing water 

 diminishes the loss of heat, and the velocity of cooling 

 changes. The ice in the transparent layers was frozen 



