388 Dr. N. Ernest Dorsey on the Surface' Tension of 



water ; then more salt is added, and so on until readings have 

 been taken on water and on solutions of (as a rule) four 

 different concentrations. Then the volume of the solution is 

 measured. I measure the volume last so that the purity of 

 the water may not be affected by any trace of grease that 

 might be on the measuring vessel. 



Since each crest is seen in two positions, measuring over 

 forty crests gives the length of twenty waves. Such a series 

 of readings as I have just described occupies about two hours ; 

 and to make sure that the surface in this length of time 

 does not become contaminated so as to affect my results, a 

 check series was taken on water. The first readings gave 

 Ti8°-9= 73'10, or Ti 8 o = 73 , 24 ; the last, about two and one-half 

 hours later, gave T 19 °. 3 = 72*91, or Ti 8 °=73 , 18. Hence the 

 results obtained from such a series of readings that extend 

 over about two hours should be not appreciably affected by 

 the contamination of the surface by the air. 



By this plan of work we obtain for each concentration two 

 sets of three independent determinations of the wave-length. 

 If the averages of these two sets agree fairly well, the surface- 

 tension is calculated from their mean. If they do not agree 

 well another series is taken, and the surface-tension is calcu- 

 lated from the weighted mean of the three series. 



Temperature Reduction. 



To reduce the results to the standard temperature, 18 de- 

 grees Centigrade, I used the formula T t — T (l— at), and took 

 a = 0*0020, which is the average of the following values : — 



Brunner 0*0019 Pogg. Ann. lxx. p. 481. 



Braun 0-0020 Wiutlemann's Handb. d. Fhys. i. p. 467. 



Frankenheim 00018 Pogg. Ann. lxxii. p. 177 (1847). 



Wolf 00018 Ann.chim. etphys. [3 J xlix. p. 269 (1857). 



Volkmann 00018 Wied. Ann. xvii. p. 353 (1882). 



Timberg 0-0022 Wied. Ann. xxx. p. 545 (1887). 



Jager . ." 00023 Acad. d. Wiss. in Wien, c. p. 245 (1891). 



Cantor 0-0021 Wied. Ann. xlvii. p. 399 (1892). 



Humphreys & Mohler 0-0020 Phys. Review, ii. p. 387 (1895). 



Sentis 0-0020 J. de Phys. [3] vi. p. 183 (1897). 



0-0020 



In this formula t is the temperature in degrees Centigrade ; 

 T t is the surface-tension at t° C. ; and T is the surface- 

 tension at 0° C. 



Since the salt solutions used were very dilute I used the 

 same temperature-coefficient (0*0020) for them ; Sentis says 



