332 Messrs. J. C. McConnel and D. A. Kidd. [June 21, 



prosecute the subject further next winter. We shall first give a 

 general account of our results, and then describe the experiments in 

 more full detail. 



Main found that a bar of ice, which had been formed in a mould,* 

 yielded slowly but continuously to tension, though kept at a tempera- 

 ture some degrees below freezing point. We began work under the 

 impression that the rate of extension depended mainly on the tem- 

 perature and tension, and that the chief difficulty lay in keeping the 

 temperature constant. But by a happy chance oar very first experi- 

 ment showed us that not merely the rate, but even the very existence 

 of the extension depended on the structure of the ice. And this is 

 a matter which seems to have been quite disregarded by previous 

 experimenters . f 



After many, and for the most part unsuccessful, attempts to obtain 

 a piece of perfectly clear ice, frozen in the mould used by Main, we 

 took a bar cut from the clear ice formed on the surface of a bath of 

 water, and froze its ends on to blocks of ice fitting the two conical 

 collars through which the tension is applied. To avoid any question 

 as to the ice giving way in the collars, where it is subjected to 

 pressure as well as tension — the bar was pierced near either end by 

 a steel needle firmly frozen in, and the measurements were taken 

 between the projecting ends of these needles. We found to our 

 astonishment that the stretching was almost m'Z, though the tension 

 was decidedly greater than that usually applied by Main. There was 

 a slight extension at first, but during the last five days the extension 

 observed was at the mean rate of only 0*00031 mm. per hour per 

 length of 10 cm., and this may well be attributed to the rise of 

 temperature which took place. The rigidity cannot have been due 

 to the cold, for during the last 24 hours the temperature was 

 between —1° and — 2°4 After the experiment, the ice was ex- 

 amined under the polariscope, and found to be a single regular 

 crystal showing the coloured rings and black cross very well. 

 The optic axis was at right angles to the length of the bar. This 

 experiment showed it was a very necessary precaution to take the 

 measurements between needles fixed in the bar itself. For whether 

 the bar extended or not, the movement of the index H (fig. 2), showed 



* The mould produced a round bar of ice 24 cm. in length and 2*8 cm. in 

 diameter, with a conical expansion at the lower end to fit into an iron collar C (fig. 2), 

 through which the tension could be applied. The other end of the bar was frozen 

 on to ice filling a similar collar B. These iron collars were faced with carefully 

 worked brass plates, and Main determined the extension by measuring the distance 

 between the plates with callipers. — July 6, 1888. 



f See Heim, ' Handbuch der Grletscherkunde,' published by Engelhorn, Stuttgart, 

 1885, p. 315. 



X We use the centigrade scale of temperature throughout. 



