153 
eral Account of the Results of Part II of the Discussion of the 
Declinometer Observations, made at the Girard College, Philadel- 
phia, between 1840 and 1845, with Special Reference to the Solar- 
diurnal Variation and its Annual Inequality.”’ The results of this 
discussion are thus given: ‘‘ The general character of the diurnal 
motion is nearly the same for the summer half year, for the winter 
half, and therefore for the whole year. The greatest eastern de- 
flection is, at a mean, reached at a quarter before eight a.M., being 
a quarter of an hour earlier in the summer, and half an hour later 
in the winter. Near this hour the declination isa minimum. The 
greatest western deflection is reached, at a mean, at a quarter after 
one o’clock p.M., a few minutes earlier in both the summer and 
winter. At this hour the declination isa maximum. ‘The diurnal 
curve presents but a single wave, slightly interrupted by a deviation 
occurring during the hours near midnight, or from ten P.M. to 
one A.M., when the magnet has a direct or westerly motion. 
Shortly after one a.m. the north end of the magnet moves easterly, 
completing the cycle and arriving at its eastern elongation shortly 
before eight a.m. This nocturnal deviation is well marked in 
winter, vanishes in summer, and is but slightly perceptible in the 
annual curve.’’* 
The second paper is an ‘‘ Abstract of a Discussion of the 
Influence of the Moon on the Declination of the Magnetic 
Needle, from the Observations made at the Girard College, Phil- 
adelphia, between the Years 1840 and 1845.’’ In the previous 
discussions of the Philadelphia observations of magnetic de- 
clination, Professor Bache had shown how the influence of magnetic 
disturbances, of the eleven-year period, of the solar diurnal 
variation and its annual inequality, of the secular change and 
of the annual variation might be severally eliminated, leaving 
residuals from which the lunar influence is to be studied. ‘‘ One 
of the first questions to determine is, how many of these residuals 
must be used to give a definite result ? and another one is whether 
numbers deduced from different parts of the series would give har- 
monious results? ‘To test both of these the observations were 
formed into three groups, one containing four thousand nine 
hundred in nineteen months of 1840 and 1841; another, six 
thousand seven hundred and fifteen results in twenty-one months 
of 1842 and 1843; and a third, ten thousand and twenty-nine 
* Proc. Amer. Assoc. Adv. Science, xiv, 74, 1860. 
PROC. AMER. PHILOS. SOC. XXXII. 143. T. PRINTED DEC. 28, 1893. 
