NATURE 



{Nov. ;, 1878 



Singapore ... . 

 Trevandrum 

 Madras ... . 

 Bombay ... . 



Pekin 



Catherinenburg . 



Ranges of Monthly Means.' 



Barometer. 



0*064 



0'II2 



o'299 

 o"a85 



o'777 

 o"3io 



Thermo- 

 meter. 



3-1 F. 



4-8 

 11-8 

 11-4 

 52-8 

 64-8 



Bar. oscilla- 

 tion for i'. 



in. 



0'021 

 0'023 



©•025 



0'025 



o'ois 



o'oo5 



Trevandrum ranges are placed beside them for com- 

 parison : — 



I have not the means for Calcutta, but they confirm the 

 relation shown for the other Indian stations. At Madras 

 and Bombay the barometric oscillation corresponding to 

 1° F., attains a maximum, it has diminished considerably 

 at Pekin, and at Catherinenburg it cannot be said to 

 exist ; for though the minimum pressure occurs near the 

 epoch of maximum temperature, the time of maximum 

 pressure may be said to extend over the seven months, 

 October to April, during which the monthly mean tempe- 

 rature varies yi" F. Also at Hobarton (42° S.) the relation 

 does not exist, though the range of monthly mean tempe- 

 rature is nearly 20°. 



When we examine the individual cases in India more 

 in detail, several facts present themselves which are 

 opposed to any relation of cause and effect between the 

 two phenomena. Thiis, as regards the epochs of maxi- 

 mum of one variation, and minimum of the other, though 

 both show considerable approximation to the solstices, 

 yet there are some marked differences. At Bombay, 

 where the barometric oscillation is greatest for a given 

 oscillation of temperature, the epoch of maximum pres- 

 sure -precedes by nearly a month that of minimum tempe- 

 rature, while the epoch of minimum pressure occurs more 

 than a month after that of maximum temperature. We 

 find also that the relation of increasing oscillation of 

 pressure with increasing oscillation of temperature which 

 holds within certain limits for different places does not 

 hold at the same place. That is to say, years of largest 

 annual variation of monthly mean temperature are not 

 years of largest variation of monthly mean pressure, and 

 ir» some cases (as at Trevandrum) the range of monthly 

 mean temperature may be nearly twice as great in one 

 year as in another.^ 



No better example, however, of the relation of the two 

 oscillations to local conditions can be given than that 

 presented by the variations at two stations within sixty 

 English miles of each other, Trevandrum within four 

 miles of the sea on the western side of the Ghats, and 

 Pallamcottah on the eastern side, nearly in the centre of 

 the burning plain of Tinneveliy, and about 13' north of 

 Trevandrum. The monsoon, which commences in the 

 end of April on the western side of the Ghats, is not felt 

 at Pallamcottah ; the annual range of temperature is thus 

 much greater than at Trevandrum. The following are 

 the mean ranges at Pallamcottah, ['derived from three 

 years' observations by Col. W. H. Horsley, of the 

 Madras Engineers, to whose kindness I owe them.^ The 



' ' From means of five years' observations at Singapore and Catherinenburg, 

 six years at Pekin, twenty-six years at Bjmbay, eight years at Trevandrum, 

 and fourteen years at Madras. 



r . * When the twenty-six years' observations at Bombay are placed in two 

 groups, one with the largest, the other with the smallest annual variation of 

 monthly rnean temperature, I find the following ranges of the monthly 

 means, derived in each case from the means of thirteen years' observations : — 



Ranges of Temperature. Pressure. 



Maximum i2-6 o"29S 



Minimum ... 9'9 0*239 



Difference 



2 7 



o"oo6 



The difference of ranges of pressure for 2°'7 at Bombay at the rate of 

 o"025 in. for 1° should be o'c67 in. The difference of pressure found is in the 

 right direction, but is not one-eleventh of what it should be if the two were 

 related as cause and effect, or even if the two were exact measures of vari- 

 ation of a single cause, 



3 The approximate height of Pallamcottah above the sea-level is 125 feet; 

 that of the Trevandrum barometer being 193 feet. 



It will be seen that the ranges at Pallamcottah are 

 more than double those at Trevandrum. The consequence 

 of this remarkable fact is that, with an equal mean pres- 

 sure at the two stations, the monthly mean pressure at 

 Pallamcottah must be 0*065 inches greater in January 

 and 0-065 inches less in June than at Trevandrum, and 

 that the oscillations appear independent of the laws of 

 equilibrium of pressure in gases.' 



That the higher pressure in January at Pallamcottah 

 does not depend on the lowness of the temperature in 

 that month is proved by the fact that at Trevandrum, 

 sixty miles distant, the mean temperature for January is 

 more than 4° F. lower, and the pressure, instead of being 

 higher on this account, is really more than half a tenth of 

 an inch of mercury loiuer than at Pallamcottah for the 

 same level. 



The question which is now particularly before us refers 

 to the yearly mean pressures. As the exact height (above 

 sea-level) and index-error of Col. Horsley's barometer are 

 unknown to me, and the direction of mean isobars can be 

 determined with any accuracy only from observations 

 with standard instruments, whose heights abov^e the sea- 

 level have been ascertained by levelling, I have employed 

 for their determination observations made at the three ^ 

 observatories of Trevandrum, Madras, and Bombay,! 

 which fulfil the requisite conditions. The following are 

 the mean pressures and mean temperatures at the 

 stations ; I have added a near approximation to the mean 

 pressure at Singapore.^ 



Barometer. Thermometer, 

 in. 



Singapore 29-904 ... 80-3 Falir. 



Tievandriam ... 29-878 ... 78*4 ,, 



Madras 29-864 ... 83-0 ,, 



Bombay 29-846 ... 79*0 ,, 



When the latitudes and longitudes given previously are 

 considered, it will be seen that the mean pressure divii- 

 nishes from the equator at the rate of 0-003 inch for each 

 degree of latitude; and that the yearly mean isobars 

 run parallel to the equator (this conclusion is independent 

 of the mean for Singapore). It will be seen also that 

 the yearly mean isotherms do not lie parallel to the 

 equator, their directions being determined chiefly by 

 local conditions. 



As the difference of latitudes of Trevandrum and 

 Pallamcottah is only 13', the difference of yearly mean 

 pressures should, by the preceding result, not exceed 

 o-ooi inches, while the mean temperatures differ by 

 nearly 7° Fahr.^ 



We may conclude, then, that, whatever may be the 

 apparent relation of the annual oscillations of pressure 



"^ I do not enter at present into the consideration of the pressure-oscilla 

 tionfrom lo A.M. to 4 p.M , which appears to show a .similar relation tj the 

 temperature-oscillation. 



- The mean pressure at Singapore is corrected by -F 0-020 m. for reduc- 

 tion to sea-level, a correction which it is believed is near the truth. The mean 

 temperatures have received no reduction to sea-level at three stations as the 

 heights are small, from 20 to 37 feet. That for Trevandrum has been 

 increased by o°-6 F. (height X93 feet) as an approximate reduction to the 

 height of the other stations. The Madras barometric mean is corrected for 

 index-error by — 0-005. See Introd. Madras Obs. 1851-55. 



3 Since by the annual oscillations 1° Fahr. is equivalent at both stations to 

 0-023 inches of mercury, if Mr. Chambers' analogi' held, the yearly mean 

 pressures should differ by 7 X 0-023 =o-i6i inches. The mean barometric 

 pressure, corrected approximately for height at Pallamcottah, is very nearly 

 the same as at Trevandrum. This is independent of the conclusions from 

 the directions of the isobars. 



