148 Royal Society : — 



nations in a detached building, appi'opriated chiefly to a close inves- 

 tigation into the direction and amount of the secular change. The 

 result is stated in the 3)d volume of the Toronto Observations, p. cxxvi, 

 and is as follows: — "The secular change of the declination from 

 1845 to 1851 inclusive was an annual increase of 1''95 of west decli- 

 nation. From July 1851 to April 1854 (two years and nine months) 

 an annual increase of 2'54 : and assuming the circumstances of anew 

 series commenced in 1855 with the same instrument placed in a new 

 building to be strictly comparable with those of the old series, the 

 increase from April 1854 to October 1855 is at the mean annual rate 

 of 3''54." The progressively increasing amount of the rate of secu- 

 lar change is a circum:<tance which, for obvious reasons, may be ex- 

 pected to follow for a time after the reversal of the direction of the 

 change. 



Attention being thus alive, particular care was taken that the 

 azimuth compass with which Lieut. Blakiston was supplied should 

 be free from instrumental error, and the practice was recommended 

 to him of repeating observations at different hours and on different 

 days. The following is a transcript of the report received from him 

 from York Fort, showing how thoroughly these directions were kept 

 in view :— ., Pechnation at York Fort, 1857. 



h m a ; 



17th August, 5 30 p.M 7 01 E. 



5 43 p.M 7 21 E. 



G 14 p.M 7 43 E. 



20th August, 5 18 p.m 7 41 E. 



5 53 p.M 7 24 E. 



26th August, 5 54 a.m 8 01 E. 



6 40 A.M 7 57 E. 



7 20 A.M 7 50 E. 



Mean 7 37 E. 



" Ten to twelve observations in each set, the compass being lifted 

 and shaken between each observation." ^ ^ 



The observations of Franklin in Se])tember 1819 gave 6 00 E. 



Those of Lefroy in July 1843 gave 9 25 E. 



Those of Blakiston in August 1857 gave 7 37 E. 



It appears therefore that the secular variation which between 1819 

 and 1843 caused an increase of east declination, caused on the con- 

 trary between 1843 and 1857 a decrease of east declination. This 

 is a reversal in the same sense as that which has been seen to have 

 taken place at Toronto. It seems probable from an inspection of the 

 intervals, and of the differences of declination-value, in the three 

 determinations above noticed, that the epoch of reversal must have 

 coincided very nearly with that of the survey of 1843-44 ; and con- 

 sequently that Lieut. -Col. Lefroy's result may show approximately 

 the maximum which the easterly declination attained at York Fort 

 before the change took place. If we might assume 1843 to be the 

 precise epoch, it is desen'ing of remark that it is the same year in 

 which the observations of the inclination at Toronto show that the 

 annual secular variation of that element changed from a decreasing 



