On the Viscosity of Ice. 



499 



The specimen broke by a jar due to the falling of a weight on the 

 floor of the room. 



On February 27th the three observations were taken, with different 

 loads on the specimen, directly after one another. "With a load of 

 5*5 kilos, (the weight of the lower collar) there was an immediate 

 extension of 0*22 mm. ; and when the load was increased to 12*5 kilos, 

 the extension at once increased by 0"04 mm. 



The three experiments, of which the results are given above, show 

 that ice subjected to tension stretches continuously by amounts which 

 evidently depend on the temperature and on the tensile stress. "When 

 the stress is great, as in No. I, and the temperature not very low, 

 there are appreciable extensions, as on February 15th, amounting to 

 as much as 1 per cent, of the whole length per day. When the tem- 

 perature is lower, and the stress is less, the extension is less, but still 

 such as can be measured. So continuous and definite is the extension, 

 that it can even be measured from hour to hour, as seen in Experi- 

 ment No. 1, when for February 15th and 16th easily measurable 

 extensions were obtained for intervals of two and three hours. The 

 quantities actually measured were generally both notably greater and 

 less than the mean, since, owing probably to inequalities in the dis- 

 tribution of the stress and to variations of texture in the ice, 

 due to internal strains produced in freezing, one side of the specimen 

 would sometimes stretch 50 per cent, more than the other. Hence 

 differential motions resulted in the ice. These motions and exten- 

 sions took place at temperatures which preclude all possibility of 

 melting and regelation, expecially in Experiments 1 and 2. In Ex- 

 periment No. 1, it was found that in three days, from February 13th 

 to February 16th, the distance between two marks on pieces of paper 

 gummed on the ice increased from 200 mm. to 203 mm., giving an 

 elongation of -J- per cent, per day. This method fails at tempera- 

 tures very near the freezing point, owing to the danger that by 

 thawing the pieces of paper may slip, but is free from this risk at 

 lower temperatures. It is in any case very rough, and is only 

 useful as a check on the other measures, since there is no question 

 here of slipping through the collars. How close the correspond- 

 ence in the results obtained is may be seen by observing that in 

 Experiment 1 there was an extension in nine days of 11 mm. in the 

 case of a cylinder of ice, which at the beginning of that time was 

 235 mm. long, and this gives an elongation of just about \ per 

 cent, per day, the same as resulted from the three days' observa- 

 tion with the marked bits of paper. 



The experiments are to be regarded rather as proving the exist- 

 ence of continuous extension under tensile stress than as deter- 

 mining its amount. The quantities observed are functions of the 

 stress, of the time between two observations, and of the time integral 



