146 ROYAL SOCIETY OF CANADA 



Traces for the latter half of December, showing the differential 

 records and the temperature at both stations, are presented in further 

 illustration of the applicability of the former to weather prognostication. 

 The temperature traces have not be^n corrected for index error, but the 

 two curves show the proper difference between the upper and lower 

 stations. 



At noon on December 17th the temperature on the mountain be- 

 came warmer than at the lower station, and so continued for ten hours, 

 the temperature at the lower point falling slightly after 7 a.m. on the 

 8th, when the change to warmer weather began and extended through 

 more than 20 degrees. 



JBefore 10 o'clock a.m. on the 18th, while the temperature at the 

 lower station was still increasing the differential record rapidly returned 

 to zero, and then showed a negative difference which continued to 

 increase at first rapidly and afterwards slowly for about 16 hours. 

 'J'he change to lower temperature at the Observatory took place 

 at about 1.30 a.m. on the 19th, and continued through 15 degrees. 

 The interval here noted between the change to a minus differ- 

 ence and the commencement of cold weather at the lower 

 station amounted to 16 hours. A diff'erence of upwards 



of ffve degrees was maintained between the stations for about 

 twelve hours, indicating steady cold at the lower station, after which tho 

 difference decreased slightly until midnight of the 19th. Shortly after- 

 wards the approach of a warm wave was noted by a return to zero and 

 the lower station showed the expected change some ten hours later (10 

 a.m. of the 20th). In this case the differential record only just reache.l 

 the zero, and after four hours fell away from it. As was to be expected, 

 the temperature change at the lower station was of correspondingly 

 small dimensions. 



Following this the record remained at or slightly below the 

 normal difference for the month, and no considerable change occurred in 

 the air temperature uiitil the 22nd at 8 a.m. when^the approaching 

 change was heralded by the differential thermometer at 3 a.m., or only 5 

 hours in advance. Again the plus indications were of slight duration, 

 and the warm period similarly brief. The following two days gave a 

 period of steady temperature with a gradual approach to colder weather 

 on the evening of the 24th, and no considerable fluctuation in difference 

 from minus 5 degrees until midnight on the 24th, when for a few hours 

 the record showed a plus sign reaching -|-3 degrees, and the air tempera- 

 ture rose to 30 degrees at noon the following day, the 25th. Similar 

 cases of advanced warning may be observed in the following days. The 

 very extensive deviation to the plus side continuing from 3 p.m. on the 



