204 



NATURE 



\^Dec. 27, 



caution of observing the extension of the bar proper by 

 measuring the distance between two needles fixed in the 

 bar near either end. We used a cathetometer in the first 

 instance, but that generally unsatisfactory instrument was 

 particularly untrustworthy in our circumstances, and the 

 small extension we found may have been due to errors of 

 reading. We applied, therefore, a system of light levers 

 to the needles, which would indicate a very minute 

 extension, though it was not w'ell adapted to measure 

 large extensions with accuracy. Under this far more 

 severe test, the bar still maintained its rigid character. 

 Between two of the readings there was a slight extension 

 of 0'044 mm. This we attributed to a sort of surface 

 crack which we found in the bar after the experiment. 

 With this trifling exception, the whole of the lengthening 

 seemed to be caused by a gradual rise of temperature 

 which took place. This supposition gave, indeed, a 

 coefficient quite concordant with the latest results ob- 

 tained by others. Even without making any allowance for 

 the rise of temperature, the mean rate of extension during 

 six days was less than o'ooo2 mm. per hour per length 

 of 10 cm., about 100 times as small as Main had found. 

 This enormous difference had nothing to do with either 

 the temperature or the tension, for the former averaged 

 about the same and the latter was slightly greater in our 

 experiment. The cause evidently was to be sought in the 

 nature of the ice itself, and we were not long in discovering 

 a satisfactory explanation. 



Ice is, as is well known, a crystalline body, and its 

 molecular structure is no doubt perfectly regular and 

 definite so far as it is revealed by the polariscope or 

 spectrometer. We have no reason to expect any bending 

 of the optic axis or gradual change of the indices ot 

 refraction within any one crystal. Every piece of ice, 

 therefore, is either itself a single uniform crystal, or is 

 built up of pieces, each of which is a single uniform 

 crystal. Thus, bars of ice fall into two classes— homo- 

 geneous and heterogeneous. Main's bars were hetero- 

 geneous, ours was homogeneous. We concluded, therefore, 

 that heterogeneous ice is plastic, while homogeneous ice is 

 rigid; and this conclusion was confirmed by subsequent 

 experiment. 



It is ger.erally impossible to tell with the naked eye 

 whether a piece of ice is heterogeneous or homogeneous. 

 But a polariscope settles the question at once. We put 

 together a rude form of polariscope in which the light 

 from a sheet of white paper is reflected at an angle of 57" 

 by a pile of three glass plates towards a Nicol prism 

 held in the same framework. We generally turned the 

 Nicol so as to make the field dark. Looking through the 

 Nicol, and holding a bar of heterogeneous ice between the 

 Nicol and the glass plates, some of the crystals would 

 look dark, some light, and some, perhaps, coloured. If 

 the crystals overlapped and interlaced much, the appear- 

 ance was very complicated ; but in any case it was easy 

 to make out the line where the interface of any two 

 crystals cut the surface of the bar. Our first bar was 

 square, with the optic axis at right angles to two of the 

 sides. It was about an inch thick, and it showed under 

 the polariscope the coloured rings and black cross of a 

 uniaxial crystal very well. And these remained stationary 

 and unbroken while the bar was moved parallel to itself 

 across the field of view, showing that it was a single 

 crystal. To obtain the ice we had put out a large bath 

 of water in, as it happened, comparatively mild weather, 

 and cut the bar from the ice formed at the top. The 

 water was from the ordinary hotel supply, the same as 

 had been used by Main. 



Glacier ice, as is well known, is markedly hetero- 

 geneous, being composed of irregular lumps accurately 

 fitting each other, each of which is a single crystal. 

 These lumps are called in German Gletscherkorner, and 

 in French grains du glacier; so in English we may 

 use the term glacier grains. They are found of all sizes. 



from that of a pea to that of a melon. But the average size 

 diminishes rapidly as we follow a glacier upwards towards 

 its source. At the surface of a glacier the ice is of course 

 quite disintegrated by the sun, and the original structure 

 has vanished, and on the side of a crevasse or in an 

 ice cave where the clear ice is seen, the grains are fre- 

 quently quite indistinguishable with the naked eye. But, 

 if a fragment of this clear ice be exposed to the sun for a 

 few minutes, the dividing surfaces of the grains come out 

 very clearly through thin films of water being formed. 

 Moreover, in each crystal a number of small disks 

 appear, perhaps the tenth of an inch in diameter, with 

 their planes at right angles to the optic axes. This 

 peculiarity helps to mark off one grain from another. 



On account of this structure it was probable that 

 glacier ice would prove to be plastic ; but it would have 

 been extremely rash to repeat the mistake into which 

 others had fallen, and deduce the properties of glacier 

 ice from experiments on other ice. Fortunately, it was 

 an easy matter to obtain access to a glacier. For the 

 restaurant at the foot of the Morteratsch Glacier and the 

 road thereto are now kept open in winter, and the dis- 

 tance from St. Moritz is only eight or nine miles. We 

 procured some pieces from the natural ice caves, whence 

 the stream issues at the foot of the glacier, and sawed them 

 into bars at our leisure. We tested three bars, which put 

 beyond a doubt the plasticity of glacier ice under tension. 

 The rate of extension varied, however, in the most extra- 

 ordinary manner in each bar, not merely with the tem- 

 perature and the tension, but also with changes in the 

 nature of the bar, due, apparently, to the process of ex- 

 tension itself. To make the results obtained with different 

 bars comparable, I shall give all the rates of extension in 

 millimetres per hour per length of 10 centimetres. The 

 first bar extended at a rate of from o"oi 3 mm. to 0-022 mm.,^ 

 the variations being attributable to changes of tempera- 

 ture. The second began at a rate of o'oi6 mm., and 

 gradually slowed down till it reached at the same temper- 

 ature a rate of 0-0029 nim., at which point it remained 

 tolerably constant, except for slight temperature fluctua- 

 tions, until the tension was increased by one-half. This 

 brought the rate at once up to o'ci 10 mm. This increased 

 rate in its turn showed a tendency to sink, more or less 

 counterbalanced by a rising temperature. This piece of 

 ice was under tension for twenty-five days, and extended 

 altogether about 3 per cent, of its length. The third 

 piece behaved in a very different manner. It began at 

 the rate of o'oi2 mm., increased its speed, with the ten- 

 sion nearly doubled, to o'026 mm., and stretched faster 

 and faster, with unaltered tension, till it reached the ex- 

 traordinary speed of r88 mm. We put on a check by 

 reducing the tension by one-third, whereupon the speed 

 fell at once to 0-35 mm., and gradually declined to- 

 0-043 mrn- The lowest temperature reached during our 

 experiments, except with the intractable bath ice, was 

 with this specimen. For twelve hours the temperature 

 never rose above - (f, and it probably averaged - io°-5. 

 The tension happened to be very light — only 1-45 kilos 

 per sq. cm. ; but the rate was easily measurable. It was 

 00065 mm. The arrangement of the grains in these bars 

 was too complex for description. The size averaged, 

 perhaps, that of a walnut. Nearly one-third part of the 

 third piece was one crystal, which ran three-quarters of 

 the length between the needles. 



Some, though not all, of the ice of the St. Moritz Lake 

 is possessed of a curious structure. It is built up of 

 vertical columns whose sections are of quite irregular 

 shapes. The thickness of each column is not quite 

 uniform ; still, the sides are nearly vertical. An average 

 column is about as thick as an ordinary pencil, and in 

 ■ length is only bounded by the depth of the clear ice— z.^. 

 a foot or more. Each column is a single crystal, and the 

 optic axes are generally nearly horizontal, though othsr- 

 wise arranged at random. The columns become visible 



