On the Lunar Atmospheric Tide at St. Helena .
By Lieut.-Colonel EDWARD SABINE , R.A. , For .
Sec. R.S. Received December 16 , 1846 , -Read January 28 , 1847 .
THE attractions of the sun and moon occasion tides in the atmosphere similar to those of the ocean .
The effects are , however , so inconsiderable in comparison with the disturbances produced in the equilibrium of the atmosphere by other causes , that hitherto observation has not succeeded in affording any clear or decisive evidence of them .
The phenomenon however has a very considerable philosophical interest as a farther exemplification of the universality of the principle of gravitation .
As the ebb and flow of the atmospheric tide must be greatest in the vicinity of the equator , and as moreover the ordinary equilibrium is less disturbed by irregular causes in that region of the globe than in the temperate zones , it was reasonable to expect that the existence of a tide in the atmosphere , depending on the position of the sun or moon relatively to the meridian of the place , would be shown , if at all , by observations made within the tropics ; and that the most favourable situation for such observations might be that of a small island surrounded by a considerable extent of ocean , and therefore comparatively free from the atmospheric disturbances occasioned by variations of the terrestrial surface ; and where also the tides of the ocean should be small .
The importance of St. Helena as a station of observation for this purpose early engaged the attention of the highly intelligent director of its magnetical and meteorological observatory , Captain LEFROY of the Royal Artillery , as may be seen from the following extract of a report received from that officer , dated June 1st , 1842 .
" 1 believe that no observations have as yet decided the question , whether any effect upon the mean barometrical pressure is produced by the moon 's daily passage of the meridian .
The existence of an atmospheric tide of this nature appeared however so interesting a subject of inquiry , and its detection so probable , owing to the extreme uniformity of the daily oscillation , that the observations of seventeen months , viz. from August 1840 to December 1841 inclusive , have been arranged with that view .
The mode adopted was this : the corrected height of the barometer ( i. e. the reading reduced to 320 FAHR. ) at the hour of observation nearest the moon 's meridian passage for every day has been entered in a central column ; and in parallel columns headed -2h , --4h , &amp;c. and +211 , +4h , &amp;c. have been entered the observations taken respectively at 2h , 411 , &amp;c. before and after the central observation .
A mean has been taken for the observations included in each lunar month .
It appears from the seventeen months thus examined , that a maximum of pressure corresponds to the moon 's passage over both the inferior and superior meridians , being slightly greater in the latter case ; and that a minimum corresponds nearly to the rising and setting , or to + 6h .
The average of the seventeen months gives the respective pressures as follows , viz.in .
Moon on the meridian. .
28'2714 Moon in the horizon ... .
282675 The difference being '0039 in .
" The latitude of the observatory is -15§ 57r ; the height above the sea , ascertained by levelling , 1764 feet to the cistern of the barometer .
Observations made in 1827 under the direction of Major-General WALKER , gave the particulars of the oceanic .
tides as follows : Rise at new moon. .
3 feet 6 in .
Rise at full moon. .
2 feet 9 } in .
Rise at the quarters. .
1 foot 5 in .
Establishment ... .
2 , 20m .
" There appears to be no establishment in the atmospheric tide , consequently the rise in the ocean will not account for the variation in the height of the barometer , because the times of maximum do not coincide .
" Early in 1842 Captain LEFROY was succeeded in the charge of the observatory at St. Helena by Captain SMYTHE of the Royal Artillery ; and from the 1 st of October of that year , the observations , which up to that period had been taken at every second hour of mean solar time , were taken at every hour , and thus in their re-arrangement in lunar time , the mean height of the barometer at the different lunar hours became better represented than had been the case under the two-hourly system .
During the first year of the hourly observations , viz. from October 1842 to September 1843 inclusive , the examination of the moon 's influence continued to be carried on at the observatory by Captain SMYTHE in a somewhat modified manner , which is thus described by him:- " The hourly observations for each day , extracted from the daybook , were grouped into lunar months , and the monthly mean for each hour found .
The observation taken each day at the hour nearest the time of the moon 's meridian transit was inserted in the centre column for that day , and the other observations disposed right and left in order .
As however the observations in this state are affected by the diurnal variation , which at St. Helena isvery regular and considerable , and by which the moon 's effect would have been overridden , the monthly mean for the hour was subtracted from each observation taken at that hour , and the remainder regarded as due to the moon 's action .
When only twenty-three observations intervened between two meridian transits ofth oon , the middle observationwas entered in both the -12 and + 12 column .
" 46 Variation of the Barometer at the several Lunar Hours , from October 1842 to September 1843 inclusive , in decimals of an inch , + signifying an excess of barometric pressure , and a defect .
1st .
From the superior to the inferior passage .
Moon on the upper +lb +2 .
+2 .
+3h +4h .
+5h .
+6h .
+7h .
+8h .
+9h .
+10 .
+11h .
+12h .
meridian .
+.0014 +'0011 +-0002 -0000 --00081 --0010 --0012 -.0018 -*0011 --0005 +'0002 +-0012 +40012 2nd .
From the inferior to the superior passage .
-12h .
-11h .
-10h .
-9h .
-8b .
-7h .
-6h1 .
-5h --4h .
-3h .
-2h .
-1 h. the upper meridian .
+'0016 +-0012 +-0009 +'0002 -0006 0 --0010 -00101 -0 --0004 +'0004 +-0003 +-0016 +'0014 We have here an average excess of barometric pressure of '0014in .
at the hour when the moon is on the meridian above or below the pole , and an average defect of ?
00115 at the period when she is six hours distant from the meridian ; making together an average difference in the height of the barometer of *00255 in .
between the hours when the moon is on the meridian and when she is six hours distant from it .
The arrangement of the observations for the investigation proving rather a heavy charge on the establishment at St. Helena , the examination was now taken up at Woolwich , and carried through a subsequent period of two years , i. e. from October 1843 to September 1845 inclusive , for which the following method was adopted : If we call b the height of the barometer at 320 at any hour of observation , and b the mean height of the barometer at the same hour during the month to which the day belongs , then b-b is a quantity which remains over after the approximate diurnal variation has been eliminated ; it is + when the barometer is higher and when it is lower than the normal state , and expresses by how mruch the barometer is higher or lower at that observation than the normal state at the same hour .
Monthly tables are then formed in which are inserted the values of bb for each hour of mean solar time .
The mean solar hours which are respectively nearest to the several lunar hours are then computed for every day , and the values of b-b are re-arranged in lunar monthly tables .
The means at the several lunar hours in each month are then taken ; and these means are finally arranged in periods of six months , yielding mean values of the barometrical variation at the several lunar hours for each half year .
These are shown in the following tables:47 1st .
From the superior to the inferior passage .
Lunar hours .
0 .
1 .
2 .
3 .
4 .
5 .
6 .
7 .
8 .
9 .
10 .
11 .
Oct. 1843 to March 1844 ... +0017 +'0012 +'0004 -'0007 --0021 --0018 -*0018 -*0016 --0019 -'0013 -0004 +-0014 April 1844 to Sept. 1844 ... +-0023 +-0020 +-0015 +-0007 -'0001 --0013 -'0028 --0027 -'0011 -'0004 -'0006 +'0009 Oct. 1844 to March 1845 ... +0021 +'0026 +'0023 +-0014 +-0007 +'0001 --0007 --0003 +-0005 +-0010 +-0017 +'0006 April 1845 to Sept. 1845 ... +0009 +'0001 -'0008 -'0018 --0021 -'0024 --0032 --0024 --0008 +-0008 +-0016 +-0026 Mean of the 1st year ... ... . .
+-0020 +'0016 +-0009 '0000 -*0011 --0016 --0023 --0022 --0015 --0009 --0005 +-0012 Mean of the 2nd year ... ... +'0015 +'0014 +'0008 --0002 --0007 --0012 -*0020 -*0014 -'0002 +-0019 +-0016 +-0016 Mean of the four half years +-00175 +-00148 +-00085 --00010 -*00090 --00135 --00212 -'00175 --00082 +-00002 +-00058 +.00138 2nd .
From the inferior to the superior passage .
Lunar hours .
12 .
13 .
14 .
15 .
16 .
17 .
18 .
19 .
20 .
21 .
22 .
23 .
Oct. 1843 to March 1844 ... +0021 +'0020 +-0015 -'0006 --0002 -'0005 -'0008 -'0007 +-0003 +-0012 +-0016 +-0018 April 1844 to Sept. 1844 ... +-0005 -'0002 --0007 --0009 -'0015 --0014 --0015 --0005 +-0003 +-0018 +-0024 +-0027 Oct. 1844 to March 1845 ... +'0011 -+'0007 -'0004 --0024 --0024 -'0026 --0026 --0023 --0018 +-0001 '0000 +-0005 April 1845 to Sept. 1845 ... +-0027 +'0022 +-0018 +-0010 --0006 --0015 -'0024 --0023 --0021 --0001 +-0006 +-0012 Mean of the 1st year ... ... ... +'0013 +'0009 +-0004 --0008 --0009 -'0010 -'0012 -'0006 +-0003 +-0015 +-0020 +-0023 Mean of the 2nd year ... ... +'0019 +0015 + 0007 --0007 -*0015 -0021 --0025 --0023 --0020 '0000 +'0003 +-0009 Mean of the four half years 1+00160 +'00118 1+00055 --00072 --00118 -00150 --00182 -00145 -'00082 -00075 +-00115 +-00155 The evidence is here of the most decisive character of a barometrical maximum at the lunar hours of 0 and 12 , and a minimum at 6 and 18 , with a corresponding progression at the intermediate hours .
If we now arrange these results in such manner that the hours are combined in which the moon is similarly situated in respect to the meridian , we have the lunar hoirary variation of the barometer as follows : Variations of barometric pressure .
Horary variation .
Moon 's , Moon 's distance From the obserFrom the obsers distance rationsm-at oher vaionsc at the distance from the At the hours following the At the hours pr eeding theationsat the aton at the om the meridian , meridian passage .
meridian passage .
hour following hours preceding Mean meridian .
the meridian the meridian passage .
passage .
hh in .
in .
h in .
in .
in .
in.n. .
in .
h01 '0 0160}+00l68 12+006168 +'00365 +-00365 +-00365 01 1+00148 } +00133 11++001 -8 00146 +'00330 + 00343 +'00336 11{ 13+'$00011 ' 23 + 001553 2+'00085 i0+'00058 2{ 14+ ' 00055 } +00070 200 ?
115 }+'00086 +-'00267 +'00283 +'00275 23 3-00010 -'00041 291-'o5 -000036 +-'00156 +'00161 +'00158 j15--.00072 J1-000 0 4-'000901 08-00058 34 16--}0011 00104 8-000800 00 70 +00093 +-00127 +'00110 4{ 5--'00135O__.s0014 5 5-00135 1--000142 -00145 -'00160 + 00055 +-00037 +'00046 56{ }-001 19--00145 6 0-1927 180-10082j 00000 00000 00000 6 18--'00182 18.'00182 48 It should here be remarked , that as the observations were made at hours of solar time , they could not of course strictly correspond in the majority of cases to the lunar hours under which the value of b-b have been arranged .
The effects of this imperfect synchronism have a tendency to compensate each other , and are probably nearly compensated , at all the lunar hours except at 0 and 12 , 6 and 18 .
These , however , are the most important hours for the determination of the maximum effect , and , supposing there to be no establishment , it is obvious that whether the observation precedes , or whether it follows the precise lunar hour to which it should correspond , the error produced will be the same in kind , viz. the maximum at O"1 and 12h will be lowered , and the minimum at 6h and 18h will be raised ; and in both cases the error will tend to diminish the apparent influence of the moon 's position on the barometric pressure .
It may therefore be presumed that the true horary variation corresponding to the lunar hours is in every case greater than the numbers which appear in the preceding table ; and that the horary variation at 011 and 1211 which should express the whole difference of pressure co ?
rresponding to the moon 's positions on the meridian and at six hours ' distance from it , is especially less than the true amount , being diminished by the causes above mentioned , both at the hours when the moon is on the meridian and when she is six hours distant .
To obviate this inconvenience and to give the results in future as much precision as they are capable of , eight additional observations , at equal lunar intervals , will hereafter be made in each day , corresponding precisely to the lunar hours of 0 , 3 , 6 , 9 ... .
21 ; and these observations will be made on Sundays as well as on other days .
We may conclude therefore as the result of the two years of observation from October 1843 to September 1845 inclusive , that the barometer at the Observatory at St. Helena is higher when the moon is on the meridian , either above or below the pole , than when she is six hours distant from the Ineridian , by an average quantity which exceeds '00365 in .
, and may be taken in round numbers as '004 in .
For the purpose of examining whether any perceptible difference in the influence of the moon on the barometric pressure could be detected at the periods of the apogee and perigee , the following method was adopted .
The epoch of the moon 's perigee or apogee being taken from the Nautical Almanac , the nearest of the lunar hours 0 , 6 , 12 , or 18 to that epoch is taken as the middle term of comparison ; if it be 0 or 12 hours , the values of b-b at the four antecedent and four subsequent lunar meridian hours ( 0 and 12 ) are taken in addition to the middle term .
to give a mnean value corresponding to the times when the mioon is on the meridian above or below the pole .
The mean of the eight intermediate values of b-b at 6 and 18 hours taken in like manner , furnishes a mnean value corresponding to the times when the moon is six hours distant from the meridian .
The difference between these mean values of b-b gives for the epoch in question , the excess of the barometrical pressure when the moon is on the meridian above the pressure when she is six hours distant from it .
Theoretically , this excess should be greater at periods of perigee than at those of apogee .
If the lunar hour nearest to the epoch of apogee or perigee be 6 or 18 instead of 0 or 12 , that hour is taken as the middle term of comparison , and the mean values of bb consist in such case of the mean of nine values at 6 and 18 hours , and of eight at 0 and 12 .
When a Sunday intervenes the same number of values of b- &amp; , viz. eight of the one and nine of the other , are taken to give the mean value in the comparison , which consequently in such case extends to a somewhat greater distance on either side of the middle term than when no such interruption occurs .
The mean excess of barometrical pressure when the moon is on the meridian , resulting from this comparison , is as follows : From 13 epochs of perigee between October 1843 and September 1844 '00407 in .
From 13 epochs of perigee between October 1844 and September 1845 '00394 in .
From 13 epochs of apogee between October 1843 and September 1844 '00341 in .
From 14 epochs of apogee between October 1844 and September 1845 '00347 in .
The number of observations from which each of the first three results is obtained is 221 ; in the fourth case it is 238 .
We have here in both years a consistent indication of the greater influence of the moon on the barometrical pressure at periods of perigee than at those of apogee .
The effect of the solar action on the atmospheric pressure is far more difficult to be shown by any mode of grouping the observations than is that of the moon , not only because it is much smaller , but because it is masked in the diurnal barometric oscillation by the much greater and variable influence of the solar heat in producing atmospheric variations .
That it is elininated , in common with the other variations which depend on the sun 's horary angle , by the process described in this paper , may be inferred from the fact that no significant difference appears in comparing the excess in the values of b--b at the lunar hours of 0 and 12 , at the periods of syzygies and quadratures , during the twelve months from October 1843 to September 1844 inclusive .
The mean values in this comparison have been taken from the observations at Oh and 12 , 6 and 18 for 36 hours preceding , and 36 hours following , the epoch of each syzygy and quadrature .
The mean excess at 0 and 12 at the periods of syzygy is 00337 , and at those of quadrature '00345 .
The quantities treated of in this communication will justly appear to many persons as extremely small ; but the consistency of the partial results , when the observations are broken into periods of six months , places beyond doubt the power of observation , in appropriate circumstances and continued for a sufficient time , to determine both the existence and the approximate amount of a diurnal systematic affection of the barometer , not exceeding four thousandths of an inch .
It also affords an instructive example of the beneficial influence of mean numerical values in the advancement of the physical sciences , and of the power which we possess through their means of progressively separating from apparently irregular phenomena , that which is constant and reducible to laws .
50

