tB.VRNEs] TEMPERATURE OF THE LACHINE RAPIDS 21 



These differences may be regarded as satisfactory, when it is con- 

 sidered under what unfavourable circumstances they were taken. 



In determining differences in temperature of the various parts of the 

 water at short intervals of time, of course all errors due to the freezing-point 

 mixture and cabin end of the thermometer were eliminated. Errors of 

 =b 005° C, and even greater, have been obtained in cases where the freez- 

 ing-point mixtures were carelessly made and not sufficiently compact, or 

 when attempts were made to use old mixtures. These either contained 

 an excess of water through melting, or were being converted into solid 

 ice on the bottom, from the fact that the floor of the shanty was of a 

 temperature much below freezing. 



To avoid errors of conduction along the thick lead tube containing 

 the connecting leads, a length of about 18 inches was buried in the snow 

 with the therraometerstem, and about a foot more was protected by build- 

 ing up snow around it. When both stems wei'e immersed at the same 

 time they were, as far as possible, buried together. From time to time, 

 while readings were being obtained, the immersed end, or in the case of a 

 zero-reading, each end was stirred, and snow repacked around it. 



To ensure more perfect uniformity, it would have been better to 

 arrange some form of helical pump, by which water could have been set 

 in circulation through the snow ai'ound the thermometer stems, by draw- 

 ing it from the bottom to the surface. A greater degree of accuracy 

 could have been thus attained, but owing to the want of space in the 

 shanty, it was thought that serious difficulties might have arisen, which 

 would outweigh the advantages to be gained. Then, again, unless the 

 other conditions in the shanty, such as its being unprotected from the 

 wind, and the proximity of magnetic bodies to the galvanometer, could 

 have been rectified, it was useless aiming at greater refinement in the 

 freezing-point mixture. 



On account of the refinement of the temperature measurements great 

 care had to be taken with the thermometer to avoid sharply bending the 

 stems or otherwise straining the fine ^^li^tinum coils. When this took 

 place during a set of readings, it was necessary to redetermine the zero 

 point. 



Time of the Observations. 



The shanty was put in place on the first day of February, 1897, and 

 on the third the instruments were taken out and observations com- 

 menced, lasting over an interval of nearly six weeks. It was impossible 

 to obtain readings every day, on account of the length of time required 

 to reach the cabin from the city. Delays occurred through the occasional 

 giving out of the apparatus, owing to the want of laboratoiy conditions. 

 College duties also prevented the writer from being away long at a time 

 during the day. Nevertheless, a great many observations have been ob- 



