314 British Association for the Advancement of Science. 



and the specific heat of the medium which encompasses the wet 

 thermometer, both of which vary, the former with the temperature, 

 and the latter with the pressure and the amount of vapour present 

 in the air. Such objection is theoretically just, and the necessary 

 corrections have therefore been investigated by the author in his 

 original paper, and may be applied if deemed necessary. Experience 

 however has satisfied him that, generally speaking, they may be neg- 

 lected, as in almost every instance their amount is considerably 

 within the inevitable errors of observation. 



The experiments instituted for the purpose of testing the formula, 

 and which are detailed in the author's second paper, were next ex- 

 plained. The principle of the first of these is as follows : if air, in 

 reference to which t, t' and t" (the dew-point) have been accurately 

 noted, be raised to any elevated temperature, and the observation be 

 repeated in the heated air as far as respects t and t\ we shall have 

 two separate sets of observations from which to calculate the point 

 of deposition ; and as the amount of moisture in the air is not altered 

 by the augmentation of temperature it has experienced, both calcu- 

 lations, provided the formula be correct, should give precisely the 

 same result, i. e. the dew-point in the first instance determined by 

 observation. Four distinct series of experiments on this plan were 

 performed by means of a very simple apparatus, and though the de- 

 pressions varied from 4 0, 7 to 28 0, 5, the calculated dew-points for 

 each series were found almost coincident, and the differences between 

 these and the observed dew-points were so trifling in amount as to 

 be clearly ascribable to unavoidable inaccuracy of observation. 



The next test experiments performed were suggested by the for- 

 d p 

 mula itself. If/" =/' ~" 07 X «n» axi ^ f" De supposed equal to 



0, a condition which can only be fulfilled in perfectly dry air, 

 f' = —X qTT, an equation from which we deduce d = 87/' x — . 



Hence by determining experimentally the depression of the wet 

 thermometer in perfectly dry air we shall be enabled to pronounce 

 upon the validity of the general method under discussion. In order 

 to observe several values of c?, air forced from a caoutchouc bag 

 was made to pass three times through about two inches of oil of 

 vitriol, and then to traverse a tube containing the dry and wet ther- 

 mometer, and the indications of these instruments were noted down 

 as soon as the latter assumed its stationary temperature. Of nine- 

 teen observations of depression thus obtained, eleven were a little 

 greater, and eight a little less than the calculated results. The mean of 

 the plus errors of the formula was "28, and of the minus errors "4 of a 



.gg .40 



degree, so that — = — "006 is the mean difference between 



experiment and calculation deducible from the whole. This sin- 

 gularly close correspondence of theory with experiment is the more 

 satisfactory because as the mean pressure for the nineteen experi- 

 ments was but a little over 30, and as the air was perfectly dry, nei- 



