NATURE 



[Nov. 25, 1880 



again to a convenient position. Operate thus, keeping the 

 surface of separation of liquid and solid at as nearly as possible 

 a constant position of 3 centimetres below the top of the tube, 

 until the surface of separation disappears. 



The temperature of tLe tube at tlie place where the surface of 

 separation was seen immediately before disappearance is the 

 critical temperature. 



It may be remarked that the changes of bulk produced by the 

 screw and mercury in Andrews' apparatus are, in the method 

 now described, produced by elevations and depressions of tem- 

 perature in the lower thermal vessel. By proper arrangements 

 these elevations and depressions of temperature may be made as 

 easily, and in some cases as rapidly, as by the turning of a screw. 

 The dispensing with all mechanism and joints, and the simplicity 

 afforded by using the substance to be experimented upon, and no 

 other substance in contact with it, in a hermetically sealed glass 

 vessel, are advantages in the method now described. It is also 

 interesting to remark that in this method we have continuity 

 through the fluid itself all at one equal pressure exceeding the 

 critical pressure, but at different temperatures in different parts, 

 varying continuously from something above the critical tempera- 

 ture at the top of the tube to a temperature below the critical 

 temperature in the lower part of the tube. 



The pressure may actually be measured by a proper appliance 

 on the outside of the lower part of the tube to measure its 

 augmentation of volume under applied pressure. If this is 10 

 be done, the lower thermal vessel must be applied, not round 

 the bottom of the tube, but round the middle portion of it, 

 leaving, as already described, 10 or 20 ems. above for observa- 

 tion of the surface of separation between liquid and vapour, 

 and leaving at the bottom of the tube 20 or 30 cms. for the 

 pressure-measuring appliance. 



This appliance would be on the same general principle as that 

 adopted by Prof. Tait in his tests of the Challenger thermome- 

 ters under great pressure (Proceedings, Royal Soc. Edin., iSSo) ; 

 a principle which I have myself used in a form of depth-gauge 

 for deep-sea soundings ; in which the pressure is measured, not 

 by the compression of air, but by the flexure or other strain 

 produced in brass or glass or other elastic solid. 



ABNORMAL VARIATIONS OF BAROMETRIC 

 PRESSURE IN THE TROPICS, AND THEIR 

 RELATION TO SUN-SPOTS, RAINFALL, AND 

 FAMINES 



T N the first part of his work on the Meteorology of the Bombay 

 ■*• Presidency, which was submitted to Government in August, 

 1S75, Mr. Charles Chambers pointed out that the variation ol^ the 

 yearly mean barometric pressure at Bombay shows a periodicity 

 nearly corresponding in duration with the decennial sun-spot 

 period (see "Meteorology of the liombay Presidency," §26, 

 p. 12), and in August, 1878, in a letter to Nature, vol. xviii. 

 p. 567, I drew special attention to this relation, pointing out that 

 the observations of the winter and summer half-years, separately 

 as well as conjointly, show that the pressure is low when the sun- 

 spot area is great, and vice veisd, but that the pressure curve lags 

 hihind the sun-spot curve. 



In November of the same year the eminent physicist, the late 

 Mr. John Allan Broun, regarding the relation thus established 

 between the variations of barometric pressure and sun-spots .-is 

 one of veiy great importance, in that it gave a probability to the 

 existence of similar laws in the variations of other meteorological 

 elements v.hich he believed was previously wanting, communi- 

 cated to the same periodical (Nature, vol. xix. p. 6) an article 

 in which he showed, from the observations recorded at Singa- 

 pore, Trevandrum, Madras, and Bombay, that the years of 

 greatest and least mean barometric pressure are probably the 

 same for all India, and from this he inferred that the relation to 

 the decennial sun-spot period found for Bombay holds for all 

 India. 



In December, 1S78, Mr. S. A. Hill supplemented and con- 

 firmed Mr. Broun's communication by giving similar data for 

 Calcutta (Nature, vol. xix. p. 432). 



In May, 1879, Mr. E. D. Archibald communicated to 

 Nature, vol. xx. p. 28, the fact (brought to his notice by Mr. 

 S. A. Hill) that at St. Petersburg the mean annual barometric 

 pressure'is high when the sun-spots are numerous, low when they 

 are few, but that the pressure epochs lag behind the sun-spot 

 epochs. 



In December of the same year Mr. Blanford presented to tie 

 Asiatic Society of Bengal a paper (Journal of the Asia;ic 

 Society of Bengal, vol. xlix. part ii. 1880, p. 70) in which 

 it was shown that the barometric obser^'ations recorded at 

 Batavia from 1866 to 1878, at Akyab, Chittagong, and Darjeehng 

 from 1867 to 187S, at Port Blair from 186S to 1878, and at 

 Singapore from 1S69 to 1878, afford more or less confirmation of 

 the results previously obtained for other stations in India." And 

 in the same paper Mr. Blanford brought forward the observations 

 recorded at the Russian observatories at Ekaterinburg, Slatoust, 

 Bogolowsk, and Barnaul from 1S47 to 1S77, and showed that at 

 the two former stations during the whole period, and at the two 

 latter during the first half of it, the barometric variations were 

 similar to those previously obtained by Mr. Hill for St. Petersburg. 

 In a subsequent letter to Nature, published in March, 1880, 

 Mr. Blanford discussed the same observations in greater detail, 

 dealing with the summer and winter observations separately, as 

 well as conjointly, and showed that the decennial variation of the 

 barometric pressure found for St. Petersburg was exliibited 

 only by the obsen-ations of the winter months. He also obtained 

 similar results for Ekaterinburg and Barnaul, but he appears to 

 have overlooked the very important facts that the range of the 

 winter curves rapidly decreases in passing from St. Petersburg, 

 tlirough Ekaterinburg to Barnauland, that the summer curves for 

 the two latter stations are, on the whole, of the same character 

 as the summer curves of the Indian stations, as may be seen by 

 comparing the dotted cm'ves for Ekaterinburg and Barnaul, 

 given in Nature, vol. xxi. p. 48, with the summer curve for 

 Bombay, given in vol. xviii. p. 568 of the same periodical. He 

 also showed that the barometric curves for Batavia, Singapore, 

 and Port Blair were, as at other Indian stations, of the same 

 character both in winter and summer. 



In 1S73 and 1874 (see British Association Reports for those 

 years) Mr. Meldrum showed that there was strong evidence of a 

 connection between sun-spots and rainfall, and he has recently 

 (ee Monthly Nolice oi the Meteorological Society of Mauritius 

 for December 1878) put this question beyond all reasonable 

 doubt by showing that the mean yearly rainfall of Great Britain, 

 the continent of Europe, America, India, and the Southern 

 Hemisphere, varies in the same way as the sun-spots, being on 

 the average great when they are numerous, small when they are 

 few. 



In my "Brief Sketch of the Meteorology of the Bombay 

 Presidency" - in 1S76, I pointed out that the abnormal variations 

 of the monthly mean barometric pressure in that year were 

 mainly variations in the intensity of the usual seasonal move- 

 ments, although at least some portion of the variations influenced 

 a wider area than the Indian monsoon region, and in the Sketch 

 for 1877 I attributed the uniformly high barometric pressure and 

 the deficient rainfall of that year to a weak development of the 

 equatorial belt of minimum pressure, probably induced by a 

 diminution of the solar heat. 



In the Report on the Meteorology of India in 1877 Mr. Eliot 

 showed that the high pressure of that year was a characteristic 

 of the whole Indian area and also of Australia. 



In my meteorological sketch for 1878 I showed that the 

 abnormal barometric movements observed at Zi-kawei in China 

 and at Manilla in 1S7S were similar to those recorded in Western 

 India ; that the latter largely influenced the rainfall of the Bom- 

 bay Presidency ; and that in former years of deficient rainfall at 

 Bombay the barometer had been relatively high, not only at 

 Bombay, but also at Mauritius and Batavia. 



In the paper (Journal of the Asiatic Society of Bengal, vol. 

 xlix. part ii., iSSo, p. 70) already quoted, Mr. Blanford has 

 confirmed the fact that the excessive pressure observed in the 

 Indian area in the years 1S76 to 1S78 extended to China and 

 Australia, and he has also shown that it affected Western 

 Siberia also. 



In my sketch for the year 1879 I have shown that these 

 uniform variations of barometric pressure are accompanied by a 

 nearly uniform variation of the percentage rainfall of all portions 



' Duringthefirst half of these periods the results for Singapore, Akyab, 

 Chittagong, and Darjeeling differ so much from each other and from the 

 remarkably accordant results obtained from the more widely separated 

 stations of Bombay, Calcutta, Port Blair, and Batavia as 

 former are of doubtful validity during the earlier years. 



= These sketches are submitted annually to Governn i 

 year following that to which they refer. See notices in Natur 

 pp. 199 and GiQ, vol. xxi. p. 384. The sketch for 1S79, containing some 

 further important conclusions with reference to the variations of rainfall and 

 b.irometric pressure, has recently been submitted to Government. 



I 



) suggest that the 

 August of the 



