10 
MR. A. W. RtrCKER AND DR. T. E. THORPE ON A MAGNETIC 
1890. In the latter year the magnet in nse at Kew was changed, but during the 
wliole of that year both the old and the new magnets were used and the comparison 
thus carried out showed that they were in accord. 
Another argument also points to the conclusion that the Kew Standard had not 
altered. 
The average annual secular change of Declination between the years 1886-91 was 
— 6'‘2 at Greenwich; 
— o''9 a.t Kew according to the Obseiwatory instruments ; 
— 5'’4 ,, ,, „ survey instruments. 
Hence the difference between the Declinations at the two Observatories altered 
during the five years 1886-91 by l''5 according to the Observatory instruments, and 
by no less than 4'‘0 according to the survey instruments. Although we are not 
prepared to deny that such a variation in the secular change is possible, yet in the 
present state of our knowledge we think that the preference should be given to that 
result which brings two such closely neighbouring stations as Greenwich and Kew 
most nearly into accord. 
Two slight modifications were introduced into our method of correcting the indi¬ 
cations of the Kew absolute instruments for diurnal variation and disturbance. 
In the first place the normal curve of diurnal variation is now determined both at 
Kew and Greenwich from the traces of the self-registering instruments on quiet 
days selected by the Astronomer Royal. We have adopted the same plan. 
Again, during our earlier survey, the corrections were supplied to us by the Kew" 
officials. The curves given by the self-registering instruments were then standardized 
by taking the mean of the differences between their indications and those of the 
absolute instruments for the whole year. Since that time the number of comparisons 
between the two, carried out in each year, has been much increased, measurements 
being now made with the absolute instruments four or five times per month. 
Hence, it is possible to get a new base line— i.e., to standardize the self-registering 
instruments—every moirth, and this plan has been adopted in our recent work. 
It was of course important to prove that this slight modification in our method 
was not the cause of the apparent change in the relations between the Kew and the 
survey instruments. Although therefore the number of absolute determinations 
made at Kew in each month was not—at the date of our 1886 survey—sufficient to 
give the base line for the month with all the certainty that could be desired, we 
repeated the calculations required to correct the 1886 observations, determining the 
monthly base line as best we could, witli the somewhat meagre data at our disposal. 
In the case of the Dips, either this change in the method of reduction or another 
referred to hereafter, or both combined, abolished the ajiparent tlifference between our 
iusti uments and that in use at Kew. This point is discussed later. 
