504 
DR. GL NEUMAYER ON THE LUNAR-DIURNAL VARIATION 
we received a mean lunar-diurnal variation curve for each half of the year, from which 
again was derived the annual curve. A more rigid examination of the results obtained 
in this way showed forthwith that irregularities in the lunar-diurnal variation mani- 
fested themselves, for which it was hardly possible to account otherwise, than that they 
were depending to some degree on the moon’s position with reference to the equator, 
whether her declination was South or North. 
After having been once convinced of the truth of this, there could not exist a doubt 
as to the mode in which the observations had to be treated. The whole series was 
arranged in groups, “ Declination of the Moon South ” and “ Declination of the Moon 
North,” rejecting all days on which the moon was close to the equator, so as to cause 
her declination to be divided between the hours of the day. The 118 groups of lunar- 
diurnal variation thus formed were subsequently classified according to whether the 
sun’s declination was South or North, which was easily accomplished, as of each group 
only the sums expressed in minutes of arc had been taken. The mean values for each 
period were derived by allowing due weight to the number of readings from which it had 
been derived, a practice which was made a rule in the course of these discussions. The 
only exception to this in the subsequent Tables occurs in those values of the lunar- 
diurnal variation for the winter and summer half-year in the single years, they being 
derived from the monthly tables by simply taking the means. This must be borne in mind 
when comparing the various values, as in some instances there are considerable discre- 
pancies caused by this difference of treatment, which are, however, quite irrelevant for 
the purpose required. It ought to be mentioned that the mean values were subsequently 
reduced to seconds of arc. 
Originally, as I have already described, I obtained yearly and half-yearly curves of 
the lunar-diurnal variation from the monthly tables ; and it is evident that the results, 
obtained by the method just described, afforded a means of checking those of the inquiry 
primarily instituted, so far as both could reasonably be expected to tally with one another, 
the principles of classification being somewhat different in the two cases. Taking this 
into consideration, the agreement between the results was such as to cause us to receive 
the subsequent mean values with great confidence. 
The values of the first Table are to be considered as derived irrespectively of the years 
during which the observations were made, and to represent mean values for the period 
from May 1858 to February 1863. We shall presently see that this point is of some 
moment, the single years differing widely in their results with regard to the lunar-diurnal 
variation of the magnetic declination. 
The values of the following Table have been thrown into curves which maybe readily 
studied by the aid of Plate XIX. ; it was considered advisable to retain the plan adopted 
in the above Table, and exhibit the nine curves in three different groups, namely : — 
