[ 503 ] 
XV. On the Lunar-diurnal Variation of the Magnetic Declination , with special regard 
to the Moons Declination. By G. Neumayer. Communicated by the President. 
Received March 11, — Read March 28, 1867. 
The hourly records of the magnetic declination, systematically kept at the Flagstaff 
Observatory at Melbourne, Victoria, during the period from the 1st of May 1858 to 
the 28th of February 1863, have been discussed with a view to determine the lunar- 
diurnal variation to which this magnetic element is subjected. The results arrived at in 
the course of this discussion, eliciting, as I believe, facts hitherto unnoticed, induced me 
to think it important to have them brought before the Royal Society, with no other 
object in view but to direct the attention of scientific men to a subject of such vast 
import for the development of the science of terrestrial magnetism. 
The process adopted in reducing the observations, in order to eliminate the solar- 
diurnal variation of the magnetic declination, is identical with the one generally adopted 
in such cases. The limit of disturbance was taken to be 3*61 minutes of arc, and all 
hourly directions which differed from their final normals by this value, or more, were 
consequently omitted from the record. This elimination of the larger disturbances 
having been effected, from every remaining reading (R) of the magnet’s direction the 
respective final normal (N) was subtracted, thus causing the residuals (R — N) to be 
devoid of the influence of the solar-diurnal variation. When the remainders are nega- 
tive, i. e. when the normal exceeds the reading, the north end of the needle is to the west 
of its mean direction, and when positive the needle deviates with its north end towards 
the east of its normal mean. The magnetic declination being east at Melbourne, we 
perceive that the negative values denote a decrease, and the positive ones an increase, 
with respect to the normal value of this magnetic element. The total number of obser- 
vations at command amounts to 38,194, of which 4,178 single readings were excluded 
from the discussion, on account of their being beyond the above-mentioned limit, and 
there remained only 34,016 readings for the purpose of determining the lunar-diurnal 
variation. Of this number 15,735 were observed in the months from April to September, 
and 18,281 during the time from October to March. 
The treatment of the residuals, with a view to classify them according to lunar hours, 
presents no particularly new features ; so much may, however, be mentioned that, prior 
to entering upon any general discussion, every month’s result was calculated separately 
and expressed in minutes. Subsequently the values for the various months were 
arranged, irrespective of the year, in two groups, viz., the sun’s declination being South 
(October to March), and the sun’s declination being North (April to September). Thus 
mdccclxvii. 3 Y 
