422 Experimental Researches on the Depressions [July, 



With exception of the four lowermost entries, the three middle (or 

 ohserved) columns of this table accord wonderfully well with the three 

 last, which are calculated by the formula above given multiplied into 

 T, (the tabular cent, dep.) ; which is variable in the first of them, 

 (that of the experiments ;) is equal to "32 for the case of humidity 

 •58 ; and is of course = for the final case, of extreme dryness. 

 Were we to suppose that the dryness of the air did not mount higher 

 than * 1 8 (second column) from some unperceived cause, the calculated 

 depressions would suit equally well from beginning to end ; and it 

 must be remembered that any disturbing force will be much more 

 felt in the low pressures. Moreover, it can hardly be expected that the 

 depression should continue to follow the same law, after the evaporat- 

 ing surface has congealed into ice. Had the ascending series of de- 

 pressions only been used, instead of the mean, the accordance would 

 have been greater towards the middle of the scale. 



It is hardly necessary to analyse any more of the present series, 

 after ascertaining that the same co-efficient is equally applicable to 

 dry and wet air. We may therefore proceed at once to the conclu- 

 sion, that the depression of the wet-bulb thermometer, ceteris paribus, 

 varies inversely as the barometric pressure, the actual variation being for 

 every case twenty -seven hundredths of the calculated variation. 



§ 4. — Depressions under augmented barometric pressure. 



It would perhaps have been better to have preceded the last enun- 

 ciation, by a description of the experiments included under this head, 

 since they obviously form part of the same series, and must be go- 

 verned by the same law. They need not detain us many minutes. 



The modification of apparatus now employed is depicted in fig. 9. 

 Between the gasometer and the brass tube furnished with the two 

 thermometers was introduced a condensed air blow-pipe ; while at 

 the other extremity near the discharge cock k', was adapted a syphon 

 barometer capable of shewing an increase of pressure up to + 12 inches. 

 By keeping up the action of the pump with the discharge cock more 

 or less open, a current of condensed air could be maintained at any 

 pressure until the readings of the wet-bulb became stationary ; for, as 

 before stated, it was upon the current only that reliance could be 

 placed ; and my endeavour was always to maintain the same rapidity 

 in the passage of the air, although small variations in this particular 

 do not, and ought not, to produce any sensible error. 



Not having used a hygrometer in this series, I trust to the de- 

 pression itself (at 30 inches) to supply the datum of the humidity ; 

 and here of course, under condensation, the moisture increases directly 



