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



The Temperature Curve. 



In all the experiments this curve is obviously of such a form that, 

 for any given value of 6, there are two t values, say ti on the ascending 

 slope, and / 2 on the descending slope ; and when has its maximum 

 value ( = <) ti = k ( = T). 



Inspection of any of the curves representing experiments in which 

 previously dry cotton was immersed, either in water or in saturated 

 air, brings out two relations which hold good with approximate 

 accuracy. The first is that the two times for any given 6 vary inversely 

 as one another, or 



tit 2 = a .............................. (I) 



a being necessarily equal to r 2 . 



The second is that 6 varies inversely as the sum of its two times, 



and the value of b is necessarily 



Selecting Experiment 20 of Series V to illustrate these rules, we 

 find that a does not vary more than about 2 per cent, on either side of 

 the value 30 '8, and that b, which is 14T2 when is at its maximum, 

 rises as falls to about 152 and then falls again towards its original 

 value. Similar, and sometimes greater, deviations from constancy are 

 shown in all the experiments. 



Equations (I) and (II) can evidently be combined to give one which 

 approximately characterises the whole curve, viz. : 



0(t + a/t) = b ....................... . (Ill), 



where t is any time, 6 has the corresponding value, and a and b must 

 have the values already assigned to them, viz., a = r 2 and b = 2r<. 

 It follows that the whole course of any such curve is definable in terms 

 of the special values assumed by r and <]>. In other words, a " law of 

 corresponding states " is arrived at, which may be most simply expressed 

 by giving to Equation (III) the form 20/< = t\r + r/t, or may be put in 

 words as follows : 



The temperature curves for all experiments, whatever the conditions, 

 should be superimposable if in each case the values are expressed as 

 fractions of <and the t values as fractions of T. 



This suggests a method for comparing the curves for water immer- 

 sion and saturated air immersion more exactly than can be done by 

 mere inspection. The results are shown in Tables VI and VII. In 

 each of these the fourth column contains the same selected values of 

 0/<, and the first column contains the corresponding actual values of 

 0. The figures in the second and third columns respectively show the 



