191 
moon’s synodic revolution about the earth by means of hourly 
barometric observations made from January the first 1886 til] 
November the fourth of the same year. This problem shows 
great analogy to that discussed in this paper, as the amplitude 
of the lunar atmospheric tide is about 0.06 millimetres and 
therefore about the same as that of the periodical variation 
of barometric daily means caused by the sun’s revolution about 
its axis. 
Here too an approximate value, 24.8 solar hours, can be 
assumed to be known by the simple observation of the fact that 
after 29 or 50 days the lagging behind of the moon with 
respect to the sun amounts to a whole revolution and, finally 
the number of records is 7440, a number somewhat less than 
that employed in the former inquiry: the only essential diffe. 
rence is this, that here the double periodical is prevalent over 
the simple periodical variation, whereas in the former case the 
latter dominates. 
From the tables in which the barometric observations are 
tabulated according to the lunar hours, the following formula, 
calculated for the ten-monthly period mentioned and reduced 
lo the epoch January the first at 1 a. m.', is derived: 
0,067 sin (Anz + 2159); 
the true value of R is 24.8412 solar hours. 
The way in which this problem has been worked out is 
exactly the same as that followed in the preceding chapters: 
If we put: 
сі ОИК 14 sin 744 Š 
“248085 Tiina AEI: 
we find for the sums of the ten gronps taken conjointly: 
AK 2810 OME + 08 o) in (onst C-+299X24.8 E+ 4.6 pa) 
sin (744 5 + 0.5 pz) 
for that of the even groups : 
sin 5 (1488 54-24) _ 
Дж 539 x 24.8£ + 5 
AK sin (1488 š + pz) о ат 4 
