1882.] Preliminary Report to the Solar Physics Committee. 411 



If this be true, if the two kinds of allied weather start together at 

 America, or nearly together, and if the magnetical moves in the same 

 direction as the meteorological weather, but only quicker than it, then 

 it is not unreasonable to suppose that the Kew magnetical weather of 

 to-day may be found to resemble the Kew meteorological weather six 

 or seven days later on. 



It is this point which I have endeavoured to test in the present 

 communication by comparing together the variations of declination- 

 range and those of temperature-range at Kew during the years 1871 

 and 1872. 



2. Bearing in mind, however, the local nature of meteorological 

 phenomena, instead of confining myself to Kew alone, I have taken 



l^the means of the daily temperature-ranges at Stonyhurst, Kew, and 

 Falmouth, these three forming the chief stations in England of the 

 Meteorological Council, to whose kindness I am indebted for a list of 

 the daily temperature-ranges at these three places for the years 1871 

 and 1872. 



These mean ranges have been compared with the corresponding' 

 Kew declination- ranges, excluding disturbances, for which I am 

 indebted to the kindness of the Kew Committee. This comparison 

 has been made after the following method : — 



3. The meteorological material, as already mentioned, consists of a 

 series of the daily temperature-ranges at Stonyhurst, Kew, and 

 Falmouth, recorded in degrees Fahrenheit and tenths of a degree. As 

 the results are merely comparative, decimal points have been omitted j 

 also, instead of taking means of the daily temperature-ranges at these 

 three places, it has only been deemed necessary to record the sums of 

 these ranges, thereby saving the division by three for each day. A 

 specimen of these sums is exhibited in Table I, column 2. Further- 

 more, in order somewhat to equalise or tone down individual fluctua- 

 tions, daily sums of four of the numbers of column 2 are recorded in 

 column 3. 



Again, as it is fluctuations of small period, say twenty-four or twenty- 

 five days, which we wish to investigate, column 4 is made to contain 

 means of twenty-five of the numbers of column 3, this mean being 

 placed opposite to that number of column 3 which has the middle or 

 thirteenth place in the series of twenty-five. Finally, the differences 

 between corresponding numbers of columns 3 and 4 are exhibited in 

 column 5, these being taken to represent the fluctuations we are in 

 search of, the sign minus denoting defect, and the sign plus excess, in 

 the observed values of column 3 with reference to the mean or normal 

 values of column 4. 



4. These various peculiarities of the method will be perceived from 

 the following table : — 



VOL. XXXIII. 



