l62 



SCIENCE. 



[Vol. XV. No. 370 



■distinctly traced in the observations made by others, who at the 

 ifclme were not aware that these results could be deduced from 

 iheir observations. 



As such, I liave selected Professor S. P. Langley's "Profes- 

 sional Papers, Signal Service, No. XV., War Depai-tment, U.S." 

 On p. 191 is a table showing the results obtained by measuring 

 the altitude between sea-level and Lone Pine, Mount Whitney. 

 Of these forty measurements, I have in the following table given 

 the ten highest (upper half of table) and the ten lowest (lower 

 half), arranged according to the height. 



Table oj Barometric Measurements of Altitude between Sea- Level 

 and Lone Pine, Mount Whitney. 



» The ten highest give an average of 80'- P. ; the ten lowest, 63° F. 



Alongside of each of these figures will be found a number rep- 

 TBsenting the weight of vapor per unit of volume contained in 

 the atmosphere at the time of observation. These latter numbers 

 are obtained by multiplying the relative humidity at Lone Pine 

 (see table in Langley, p. 177) by the elastic force of vapor 

 (Glaisher's tables) . 



The difference between the maximum and minimum result is 

 .330 feet, or 14 per cent of the ti-igonometrically surveyed height, 

 3,760 feet. This latter number was obtained by the engineers 

 who built the railroad passing Lone Pine. Langley's party 

 •wrote to the engineers for this information, and awaited their 

 leply with considerable anxiety. 



From the table it will be noticed that the amount of vapor in 

 the atmosphere was considerably less when the ten highest re- 

 sults were obtained than when the ten lowest were obtained; 

 and, as perhaps may be better illustrated by the accompanying 

 diagram (Fig. 1) , tliere appears to be an unmistakable relation 

 between the measured heights and the humidity of the atmos- 

 phere. The heights are here placed at distances from the verti- 

 cal line to the left proportional to the amount of vapor in the 

 air, and the line a b shows the general tendency of the figures. 

 That these do not follow the line a b more closely, may be ac- 



counted for by the humidity of the air having been measured 

 only at one end of the air-column ; namely, at Lone Pine. This 

 seems to prove the greater buoyancy of cold vapors compared 

 with warm. 



We may, however, trace another coincidence in our table. It 

 will be noticed that the ten highest results are generally from 

 observations taken at noon, when the thermometers showed a 

 relatively high temperature, while the ten lowest results are all 

 (except one) from observations taken at 9 p. m. , when the tem- 

 perature was considerably less This is only what we might 

 have expected, and shows that the buoyancy or tending upwards 

 of the vapors in the atmosphere is considerably greater at a lower 

 temperature than at a higher, as explained elsewhere. 



Professor Langley' s observations give results corresponding to 

 those 1 obtained at Rouen, and the relations here pointed out 

 may be traced in numerous works from ancient and modem times, 

 though perhaps not in all. However, it should be remembered 

 that it is not so much my present purpose to show liow the meas- 

 ui-ement of altitudes by means of barometers may be carried out 

 with gi-eater accuracy than hitherto (this method being highly 

 unsatisfactory for obvious reasons) as to show that the hidden 

 agencies which are at work in the atmosphere, and without as- 

 suming which the whole atmospheric problem remains unsolved, 

 may be distinctly traced in the observations carried out by others. 

 Surely we must expect to find the secrets well concealed, or they 

 would have been demonstrated ages ago ; but here, as elsewhere, 

 it is the instances when ' 'the sky is unobscured by clouds, ' ' to 

 use a figure of speech, which we must select to make our ob- 

 servations, and Langley's table is such an instance 



The results of my experiments above referred to showed that 

 an air-column 150 metres high, between the top and base of this 

 tower, became 1.3 per cent lighter by an increase of atmospheric 

 humidity, indicated by an increase of elastic force of vapor of 

 from . 2 to . 3 of an inch pressure. The temperature wa's reduced 

 to 40° F. , and the atmospheric pressure to 30 inches. 



According to Glaisher's "Hygrometrical Tables," one cubic 

 foot of dry air at 40° F. at a pressure of 30 inches, weighs 

 557. 8 grains, while one cubic foot of saturated air weiglis 556 

 grains. The difference, 1.8 grains, is about .0033 per cent of 

 the whole weight. The dry air, by becoming saturated, has 

 therefore suffered a loss in specific gravity of .0033 per cent, or, 

 what is pretty nearly the same, it has been expanded . 0033 per 

 cent. 



According to my observations at Rouen, the loss in weight 

 would, under similar conditions, have been 3. 3 per cent, or ten 

 times greater than shown by the tables. 



How are these seemingly contradictory results to agree ? The 

 method of taking the weight of a certain volume of air confined 

 in a vessel, by which the tables have been computed, is emi- 

 nently adapted to give us the exact specific gravity ; and the ex- 

 periments have been repeated so often by excellent obsei'vers that 

 we have no reason to doubt their correctness. If, therefore, we 

 take it for granted that only one-tenth of the loss in weight sus- 

 tained by the open column of air was due to expansion, the rest, 

 or nine-tenths, must have been due to the buoyancy' of the 

 aqueous vapors, which would carry a part of the weight of the air- 

 column ; and this force could under no circumstances have shown 

 itself under the experiments with air in confined vessels, whose 

 absolute weight is taken in a vacuum. 



We have hereby been able to demonstrate the buoyancy of 

 aqueous vapor in the atmosphere as a force that must influence 

 the readings of the barometer very considerably, and we now un- 

 derstand fully why the readings of the barometer are lower when 

 the atmosphere is moist than when it is dry, as in an anti- 

 cyclone or air-cushion, simply because the gi-eater amount of 

 vapoi-s in the moist air carries a greater portion of the weight of 

 the air-column overhead than when the air is in a dryer state. 



The atmosphere being a perfect mechanical mixture of air 

 (O-t-N + COo) and aqueous vapor (H^O) , the buoyancy of the 

 latter must mainly depend upon the difference in specific gravity 

 between the vapor and the air by which it is surrounded ; and 



1 See my paper, '• On the Cause of the Diurnal Oscillation of the Barom- 

 eter," in Engineering, London, Jan. 11, 1889. 



