April 14, 1881] 



NA TURE 



557 



January .. 

 July 



35-0 Fahr. 

 74"i .. 



9'690 

 9-105 



Now as the difference of the mean temperature of these 

 two months is nearly 40^, according to what has been 

 said before, the expansion of the air from January to 

 July should have increased the pressure at the upper 

 station by about o'S inch (eight-tenths of an inch of 

 mercury) and have diminished the difference of pressures 

 by that amount. The observed diminution is, however, 

 less than o'6 inch, a deficiency of two-tenths, which has 

 to be accounted for in some other way. 



In the case of the annual variation of atmospheric 

 pressure near the sea-level, we have to consider only the 

 atmosphere .above the station ; but m that for a high 

 station we have a much more complicated problem in 

 which the atmosphere below the station plays an im- 

 portant part, a part also which may vary greatly, according 

 as the upper station is on a mountain top, on a gradually 

 ascending height, or on high table land. 



It was for these reasons that in the article (N.vrURE, 

 vol. xix. p. 6) under consideration, nothing was said of 

 the annual variation on mountains or at great heights ; 

 the results were obtained from observations at stations 

 little above the sea level, and from tiiem it appeared that 

 for every degree Fahrenheit on the range of monthly 

 mean temperature, there was a variation of from '021 

 inch at Singapore, to 025 inch at Bombay and Madras, 

 in the range of the monthly mean barometric height ; 

 while at Pekin the amount was diminished to '015 inch. 

 The conclusion that this ratio was nearly constant in 

 India referred necessarily to stations under similar condi- 

 tions. Mr. Archibald's letter has brought to notice a 

 series of Indian observations made at stations farther 

 from the sea and in North India, which are given in 

 an interesting and useful work by Mr. H. ¥. Blanford, 

 Meteorological Reporter to the Government of India, 

 and which merit consideration. The means employed 

 by me previously were derived from observations at 

 standard observatories during five to twenty-six years; 

 and it was mentioned that, as single years gave somewhat 

 variable results, the monthly means from several years' 

 observations were necessary for accurate values. In the 

 cases to be found in Mr. iilanford's work, several are of 

 observations during only two years for even lesS;, and we 

 have no exact idea of the weight which in any case can 

 be accorded to the observations at the stations. As, 

 however, the results for different stations in general con- 

 firm each other (the exceptional cases will be noticed 

 apart), I shall here form a series of groups of stations, 

 each group under nearly the same local conditions as 

 regards latitude, height above, and distance from, the 

 sea. The value of /•, the variation in thousandths of an 

 inch, of presSure-range for I' F. of the temperature- 

 range, is the same for all stations in each group, with the 

 exceptions referred to below. 



Thus the most northerly group near latitude 30' \. and 

 600 miles from the sea, viz., Lahore, Delhi, Dera Ismail 

 Khan, Mooltan, and Roorkee, gives the following result : — 



tude of 

 group (X). 



304 



Mean height Mean 

 above the annual tem- 

 sea (//). perature(/). 



75-2 



Mean ranges of Mean baro- 



temperatute and metric 



pressure oscillation 



I for 1° Fahr. 

 I I 



(-:-o- 



39'o o'62o 



This group extends about 450 miles from Delhi to Dera 

 Ismail Khan, and the heights vary from 440 to 890 feet. 

 The value oi k is the some for each station. 



The next group comprises Bareilly, Lucknow, Jhansi, 

 and Benares, for which the following are the values of 

 the different elements : — 



, feet. , , in. 



26 520 777 32-4 0-554 0-017 



The distance of the e.xtreme stations is not 300 miles ; 

 the heights vary from 270 feet at Benares to 860 feet at 

 Jhansi. For each station k — 0-017.' 



The third group includes Patua and Gya, and the 

 values are — 



25! 260 7S-5 28-0 0-537 0-019 



The distance between the two stations is about 70 miles ; 

 the heights are 179 and 347 feet respectively, and the 

 values of k are within half a thousandth of each other. 



The fourth group includes Berhampore, Burdwan, 

 Dacca, Goalpara, Chittagong, Calcutta, and Saugor 

 Island, for which the values are — 



\ h t i^t up k 



feet. , . m. 



23J 100 78-0 18-7 0-471 0-025 



With the exception of Goalpara and Chittagong, these 

 stations all lie in the Gangetic delta, at nearly the same 

 elevation, and with an extreme variation in the value of 

 k amounting to 0-004 inch.- 



The fifth group includes Bombay alone, the values for 

 which are — 



X 7; t A< A/ * 



feet. , , in. 



iS^ 37 78-8 1 1 -6 0*287 0-024 

 The sixth, Madras, Negapatam, Trichinopoly, and 

 Madura, with the values — 



,\ h f ^'t Ap h 



feet. . , in. 



115 190 S1-7 in 0-243 0-021 

 The seventh, Colombo, with the values — 



h 



^t 



feet. 

 6i 42 81-3 3-6 0-069 0-019 



It will be seen from this table that the value of i- is a 

 maximum '= 0-025) nearly under the tropic of Cancer, 

 that it diminishes thence both north and south. This 

 diminution cannot, however, be attributed altogether at 

 least to the latitude, since the value is nearly the same at 

 Pekin and at Delhi. 



The variation of the value in Northern India from the 

 maximum near Calcutta has an apparent relation to the 

 height above the sea and the distance from it,^ as well as 

 to the amount of the temperature oscillation ; the value 

 of k diminishing as the others increase. We cannot, 

 however, relate the value to any one of the other 

 variables. No doubt height above the sea may have 

 some influence, as has already been shown, but here 

 mountain stations have been avoided ; and while in the 

 Lahore group the heights vary from 420 feet (Mooltan) 

 to 890 feet (Roorkee), the value of k remains constant. 



It should be observed that the temperature ranges are 

 those of the stratum of air a few feet above the soil, while 

 the ranges of pressure include the whole atmosphere 

 above each station. We see that the value of k dimi- 



• There appears to be some difference in the values of i/ as got by Mr. 

 Broun and those derived from my edition of the " Vade Mecum " for these 

 two groups. Thus from my edition (1S77), the value of A/ for Group I. is 

 o'6i4, and therefore * =0013 ; there are also slight differences in the value 

 of k at each station. In Group II. * is 0017 only for Bareilly and Lucknow, 

 the value fjr Jhansi and Benares being ooi6.~E D. A. 



^ The figures for the fourth group have been supplied, and the fifth, sixth, 

 and seventh groups have been added as the text seemed to imply their 

 intended formation by the author. — E. D. A. 



3 Mr. Archibald has considered height not to be an element, and he cites 

 in evidence Luclcnow and Sib.-agar. very nearly at the same height ; the 

 former having * = o 017 and the latter k = 0-028 (N.iTl-RE, vol. xx., foot- 

 note p. 54 . T.le latter value, hoivever, is a mistake, for at Sibsagar k = 

 0018 verj- nearly as at Lucknow. Mr. Archibald's conclusion is that it is 

 not height, but distance fir. m the sea, as the stations are at widely different 

 distances from the sea. We should, owing to this error, be obhged to 

 reverse the conclusion, and say it is not distance from the sea, but height, 

 that is in question ; but, in reality. Sibsagar though not so far from the sea, 

 is still as far as Patua or Gya, with nearly the same value of k. 



[•see my note on this point, Xatlke, vol. xxi. p. 131. — E. D. A.] 



