190 
DR. C. CHREE: DISCUSSION OF KEW MAGNETIC DATA 
The H optical arrangements are the best, and it suffered least. The erection in 1892 
of a new upper story to the Observatory, with iron girders, produced irregularities 
on several days, which could not be satisfactorily dealt with. There were various 
other discontinuities associated with movement of iron in the building, or with 
changes of sensitiveness in the vertical force magnetograph, which rendered the 
omission of certain days expedient. But, everything considered, the number of days’ 
trace which could not be utilised was wonderfully small, a fact reflecting credit on 
the staff, especially Mr. T. W. Bakeb, who had charge of the magnetic instruments 
during the whole period concerned. It was judged important to have a complete set 
of values of the absolute daily range (maximum less minimum) during the period. 
The loss of a good many hours’ trace necessarily introduces some uncertainty into the 
daily range, because one at least of the extreme values might fall during the time 
lost; but in many cases one can be reasonably sure that the range deduced from the 
part of the trace that is complete is the full range, or at least very approximately so. 
In all doubtful cases recourse was had to the corresponding Falmouth curves, which 
were kindly lent by Mr. E. Kitto, then Superintendent of Falmouth Observatory. 
In a few cases, while the ranges accepted were derived essentially from the Kew 
curves, a small correction was applied which was based on a comparison of the Kew 
and Falmouth curves. In a few other instances the range accepted was derived from 
the Falmouth curve alone. Experience showed that the agreement between Kew 
and Falmouth ranges was usually so close that the uncertainty thus introduced into 
monthly or similar mean values must be wholly negligible. Finally, there were two 
or three cases in which one of the traces during a magnetic storm had gone beyond 
the limits of registration at the same time both at Kew and Falmouth, or the latter 
trace was otherwise incomplete. In such a case there was nothing for it but to take 
the edge of the sheet as representing one of the extreme values. On one occasion, 
February 14, 1892, the estimate thus made of the D range was not impossibly a very 
appreciable underestimate, though hardly to the extent of exercising an effect of 
more than a few tenths of a minute in the monthly mean. 
The number of days omitted from the ordinary day diurnal inequalities through 
imperfections of the trace was so small that we shall not be far wrong if we disregard 
them when comparing different months or years in Tables I. and II. The curves D, 
H or Y, of the same year were dealt with at one time, while a considerable interval 
sometimes intervened between the consideration of two successive years. Thus 
Table II. is probably a more reliable index than Table I. to the fluctuation of dis¬ 
turbance in individual elements. Since, however, the D, H and Y curves were 
considered at widely different times, when their indications in Table I. agree the 
result may be accepted with some confidence. The two quietest years were 
undoubtedly 1890 and 1900, the years of lowest sunspot frequency. The last seven 
months of 1900 contained no single day considered disturbed for any of the elements, 
and may be accepted as the quietest period of the 11 years. The year of largest 
