248 



Dr. 0. Masson. On the Wetting of [Apr. 25, 



any series of tests performed under similar conditions, k is from the 

 very nature of the case a variable ; for it must increase gradually in 

 proportion to the vapour absorbed. As a fact, however, it will be 

 shown that the error arising from a neglect of this increase of k in the 

 application of Equation (V) is negligible, provided that its value near 

 the beginning of the experiment be fairly correctly estimated ; for it is 

 only at this end that k6 is at all large in comparison with cA, and by 

 the time that k has undergone any noteworthy increase JcO itself may 

 be regarded as a mere correction. 



The values of c and k may be deduced from the experiments them- 

 selves in the following manner: In the first place, since at the 

 turning point (when 6 = <j> and t = T) dOjdt = 0, it follows from 

 Equation (IV) that here cd = dm/dt. This value of dmjdt can easily 

 be got with fair accuracy by measuring the slope of the m curve at 

 the time T : hence, c is approximately determined. But a more exact 

 value can be obtained by measuring A for a long experiment in which 

 6 has become a mere fraction of 1 and A is very large, so that kO 

 may be fairly neglected in Equation (V). The value of c being fixed, 

 that of k (taken as constant) follows by applying this equation to 

 experiments of short duration. The following case serves for illustra- 

 tion : The m curve for Series V, in which r 5*55 and <j> = 12*72, 

 gave about 3*1 as the value of dm/dt at 5 '55 minutes ; hence, 



3*1 

 c ~ jTpw = 0*25 approximately. Experiment 20 of the series showed 



m = 187*9 in 720 minutes, when was only 0*22 and A was 723*3 ; 



187*9 

 hence c = >TOQ.Q = 260. This last value was found to give very 



consistent results in the other experiments if k were taken as having 

 the value 1*1. The figures in the column headed m (calcd.) in Table V 

 were got from these values of c and k. In the case of Series IV 

 (Table III), where a different thermometer and less cotton were used, 

 and where the atmosphere of the bath (saturated at 18 instead of 25) 

 was less moist, slightly different "figures were required, viz., c = 0*25 

 andfc = 1*0 

 If the curve were rigorously defined by Equation (III), it would 



follow that A = ^ log e and 6 = , so that m could be calcu- 



lated from the time by the equation 



which involves the three constants, a ( = r 2 of the curve), J&e, and' 

 kb. As a matter of fact, a very fair approximation to the actual 

 m values may be got in this way, though the agreement is not quite 

 so good as when the measured A and are used, in which latter 



