OP ARTS AND SCIENCES. 21 



ferences in rj^ : rj^ of 1, 2, etc. pev cent from my results, and afford a 

 convenient means of comparing the relative concordance between the 

 results of various observers. It should be borne in mind, however, 

 that some of the lines are plotted from means, some from single ob- 

 servations, and others from equations deduced either by the observers 

 or by myself. The individual points, therefore, must not be used with- 

 out qualification as a measure of the precision of the results obtained 

 by any single observer. 



Carbonic Acid, COj. 



The curve OA on Plate I. is from the equation deduced at page 17 

 for my own results. There can be no question as to the decided cur- 

 vature shown being far beyond the limits of variable experimental 

 error. My computations have also shown that the third power of 

 the temperature in the linear equation was necessary. The experi- 

 mental means are marked along the curve. 



The curve Z> OB is from an equation of the same form 



in which A and B are computed from the results of A. von Ober- 

 mayer * with capillary " I." The term Ct^ was omitted because of the 

 smaller number of points and their somewhat less precision. That 

 the line represents approximately the data will be seen from noting 

 the distribution about it of the observed points. The curvature is here 

 also undoubted, although Obermayer merely remai-ks that " it follows 

 from these experiments with some probability that the increase in 

 coefficient of friction with rising temperature is less at higher than at 

 lower temperatures." A comparison of this curve with mine shows 

 that Obermayer's results and my own are in substantial agreement as 

 to the rate of change in b, the curvature of the two lines being very 

 nearly the same. Indeed, if my results be thrown into the same equa- 

 tion, using only the first and second powers of t, the value of B becomes 

 0.00000136, while that from Obermayer's data is ^—0.00000137. 

 The value of A from my results is, however, considerably larger than 

 from Obermayer's, so that his value of rjf : j;,^ at 100° is about 0.7 per 

 cent below mine, and at 200"^ about 1 per cent below. The line OC 

 is from the equation 



r,, = rj„ (1 4- 0.003582 t — 0.0000010.5 t-), 



* A. voM Obermayer, Wicn. Ber., Ixxiii. 4G8 (1876). 



