Dec. 27, 1 888] 



NATURE 



205 



to the naked eye when the ice begins to melt, and, if this 

 melting is caused by sunshine, they often become quite 

 detached and fall apart. The appearance presented on 

 the lake when the ice melts in the spring is described as 

 very curious. The crackling of the breaking columns, 

 when the loose ice drifts against the shore, can be heard 

 at some distance. It would be interesting to learn if 

 such columns have been noticed in England. (Prof 

 Heim informs me that he has found a columnar structure 

 in lake ice in the Swiss lowlands, but the optic axes were 

 all vertical.) A few experiments we made on freezing 

 water in a bath led us to attribute this structure to the 

 first layer of ice having been formed rapidly — for example, 

 in air below — 6° C. No doubt the nature of the first 

 crystals formed settles the structure of all the rest of the 

 ice. 



This lake ice afforded a capital opportunity for testing 

 our notion that the crystals themselves are rigid, and 

 that the apparent plasticity is due to some action at the 

 interfaces of the different crystals. We first tried a bar 

 whose length was parallel to the columns. This was, 

 really, trying to stretch a bundle of long thin crystals. 

 We were able to measure an extension, but it was ex- 

 cessively small, amounting to about o"i2 mm. on one 

 side of the bar and 007 mm. on the other during 208 

 hours, giving a mean rate per hour per length of 10 cm. 

 of 000046 mm. I do not believe that the crystals 

 stretched by even this small amount. For those that 

 were slightly inclined to the direction of pull would be 

 pressed against their neighbours, there would be yielding 

 at the interfaces, and consequent minute lengthening of 

 the bar. We next cut a bar such that tl)e columns ran in 

 a slanting direction across it at an angle of about 45° to 

 the length. The difference was very striking. The new 

 bar stretched at a rate of o"oi5 mm. per hour per length 

 of 10 cm., — more than thirty times as fast. 



Towards the end of the winter we determined to try 

 the effect of pressure, and after some thought decided on 

 the following arrangement, which proved in practice very 

 satisfactory. We found in Dr. Main's stock two sheets of 

 thick plate glass, about 25 centimetres by 17. We laid 

 one of these on the table, on it three pieces of ice, and 

 on them the other glass plate. The three pieces were cut 

 as nearly alike as possible, each being about an inch cube. 

 So they were short and thick enough to preclude the 

 likelihood of bending. They were arranged at the angles 

 of an equilateral triangle 9 cm. in the side. Pressure was 

 applied by means of a lever and weight at a point verti- 

 cally over the centre of this triangle, so the pressure on 

 each block of ice was the same. Measurements of the 

 distance between the plates were taken with calipers at 

 three points at the edge, so selected that it was easy to 

 calculate from the measurements the contraction of each 

 block. Our first experimental result was that the coefficient 

 of friction of ice on glass is very small. The moment 

 the weight was applied, the three pieces of ice shot out on 

 to the floor. Afterwards this inconvenient tendency was 

 held in check by freezing pieces of paper on to the ends 

 of the blocks. 



Three pieces of glacier ice showed that this substance 

 is just as amenable to pressure as to tension. The mean 

 rates during five days were respectively o'035 mm., o'o56 

 mm., and 0-007 mm. per hour per length of 10 cm. We 

 could not discover any material difference between the 

 three under the polariscope. They were all composed of 

 smallish grains averaging perhaps 7 mm. in diameter, and 

 all three were from the same lump. They were under 

 exactly the same conditions of temperature, and under, at 

 any rate nearly, the same pressure, and yet the second 

 piece gave eight times as much as the third. Of course 

 the arrangement of the interfaces was very complicated 

 in both pieces, and it may have been much less favourable 

 to distortion in the third, but it seems more probable that 

 there was some obscure difference in the state of the 



interfaces. Bubbles, at any rate, seem to have had no 

 bearing on the matter, for the third piece contained far 

 the most, and the first piece the fewest. 



We next tried lake ice with the columns vertical. The 

 mean rate of the three pieces during four days was 0001 

 mm. per hour per length of 10 cm. This was only just 

 perceptible to the calipers, and we think it may have 

 been entirely due to the yielding of the films by which 

 the paper was attached or to the same cause as in the case 

 of tension. 



Our evidence for the rigidity of an ice crystal rests on 

 three experiments. One of these was on a single crystal 

 of the bath ice, and tension was applied ; and the other two 

 on lake ice with the stress applied parallel to the columns : 

 tension in the first case, pressure in the second. These 

 showed that the plasticity of an ice crystal is either non- 

 existent, or is at any rate of a very different order of 

 magnitude from that of ordinary heterogeneous ice. The 

 optic axis in the first case was exactly at right angles to the 

 stress, and in the two latter it was not very far from that 

 position. It would have been perhaps more satisfactory 

 if we had applied stress in other directions. But it seems, 

 a priori, very unlikely that any homogeneous substance 

 should be rigid in one direction and plastic in another, 

 and in our Royal Society paper we have given more con- 

 clusive reasoning to show that the rigidity must extend to 

 the direction parallel to the axis. 



If a bar composed of a number of crystals of irregular 

 shape stretches, while remaining compact, the crystals 

 must necessarily change their shape. It is probable, 

 therefore, that molecules separate themselves from one 

 crystal, and moving across the interface attach themselves 

 to another. But to unravel the laws which govern the 

 direction and rate of the motion of the molecules 

 further experiment is necessary. Mr. Buchanan's experi- 

 ments, recently described in Nature (vol. xxxv. p. 608^ 

 xxxvi. p. 9), throw some light on the matter. They 

 render it likely that a large part of the soluble impurities 

 in the ice will be collected at the interfaces, and will keep 

 a certain amount of water in the liquid state. This liquid^ 

 however, must be a very thin film, for it does not interrupt 

 the optical continuity. If the thickness of the film were 

 not small compared with a mean wave-length of light, 

 there would be reflection, and the interface would be 

 visible to the naked eye. Nevertheless an invisibly thin 

 film might play a very important part in providing a 

 mobile medium for the transmission of the molecules. 

 According to Mr. Buchanan, the amount of liquid present 

 would be roughly inversely proportional to the nupiber 

 of degrees below 0° C. This law is very accurate near 

 0° C. With any one salt the amount of liquid at low tem- 

 peratures would be rather greater than is given by the 

 law, but at a certain temperature, the freezing-point of 

 the cryohydrate of that salt, the liquid would completely 

 solidify. According to Guthrie, the cryohydrate of CaCl2 

 freezes at - ZT C, of NaCl at - 22^, of Na.,S04 at as high 

 a point as - 0-7. If this thin film of liquid' be an essentia 

 factor, ice should be perfectly rigid at a temperature low 

 enough to freeze all the cryohydrates. On the other hand, 

 the amount of liquid should become indefinitely great as 

 zero is approached, so that the plasticity might be expected 

 to be very largely increased when the air surrounding the 

 ice rises above zero. We did not find this was the case. 

 In a tension experiment on an icicle, the surrounding air 

 for five hours was at about -|- o°'5 C, and yet the rate of 

 extension was not strikingly greater than it had been a 

 few degrees lower. 



The temperature variations proper were so small com- 

 pared with the irregular variations spoken of above, that 

 it was difficult to secure any satisfactory measure of them. 

 Still, I have a few figures to offer. In the case of the 

 second piece of glacier ice, while at - 3'*5 the rate was 

 0*0029 nim., two days before and two days afterwards it 

 was about 00020 at -5', and a few days earlier o'ooi3 



