100 Influence of the Moon on Declination of the Magnetic Needle. — 
we moon ~ no specific constant action of deflection on the © 
edle. coefficient of the first term is small, the character — 
ob 
values are indicated by dots and the smooth curve results from 
the equation. 
The difference between the curves of deflection for the eastern 
and western hour angles shown in the curve, enables us to de 
termine the diurnal lunar tide and is drawn on the plate and 
represented by the first term of the formula. The interference — 
of the two curves gives the observed form. 
The curves all agree in their distinctive characters, and show 
two east and two west deflections in a lunar day, the maxima 
W..and E. occurring about the upper and lower culminations 
and the minima at the intermediate six hours. The total range 
hardly reaches 05. These results agree generally with those 
obtained for Toronto and Prague. From 8000 to 10000 obser- 
vations seem to be required to bring out the results : aatistactorilf 
and the best results are derived from the use of both groups. 
ondary maximum occurs fourteen minutes after the upper cul- 
mination and amounts to 018. The principal maximum occurs 
at 64 17™ after the lower culmination, the easterly deflection be 
ing 0'22. The secondary maximum at 64 03™ after the upper 
culmination has a deflection of 0-19. The greatest range is 27” 
and the secondary 22”. The epochs of the maxima and minima 
are found from the formule to be at a mean, ten minutes after 
culmination. The probable error of a single computed value of — 
the lunar diurnal variation is +132. The Toronto observations — 
ve +1'"37 from more than twice the number of observations, s° 
that ris 8 observations appear to be worthy of every 
confi to from the second investigation, embracing — 
vey 14, 000 observations, the western and eastern deflections bal- 
anced, giving for the range 88’"3. The e observations also 
confirm the nearly equal deflections paver to the west and pe 4 
The epochs of the maxima and minima were found from the four — 
roots of the equation 0=0-029 cos (6+295°) +0: 414 cos (20-4859) 4 
which gave ten minutes as the mean time elapsed between the — 
tion. If we take the four phases into account the lunar action — 
seems to be retarded ten minutes, which may be termed the lunar — 
