Intelligence and Miscellaneous Articles. 



73 



temperatures (the above hold for 20°), and the following result was 

 obtained : — 



0°. 



20°. 



40°. 



60°. 



80°. 



n. 



X. 



n. 



X. 



n. 



X. 



n. 



X. 



n. 



X. 



8-418 



2-848 



8415 



2 764 



8-411 



2-701 



8-405 



2-656 



8394 



2-628 



16-224 



1-384 



16-215 



1-328 















36-695 



0-715 



36-677 



0-675 



36-653 



0-645 



36-617 



0-623 



36-551 



0-608 



37160 



0-709 



37140 



0-670 















64-668 



0-477 



64-635 



0450 



64-591 



0-430 



64-525 



0-415 



64-404 



0-404 



It is apparent from this that the want of agreement between 

 theory and experiment is to be sought in the dependence of surface- 

 tension on temperature on the one hand, and on the wave-length 

 on the other. If this dependence is eliminated, we get the following 

 values : — 



I , 



T 



X. 



T 



J -20- 



X. 



T 40 . 



X. 



T 



■ L 60' 



X. 



2-628 

 0-608 

 0-404 



T 



00324 

 0-0393 

 0-0402 



2-848 

 0-715 

 0-477 



00600 

 00669 

 0-0678 



2-764 

 0-675 

 0-450 



00490 

 0-0557 

 00566 



2-701 

 0-645 

 0-430 



00413 

 0-0479 

 00491 



2-656 

 0-623 

 0-415 



00360 

 0-0429 

 0-0439 



From this we obtain for 0° the formula 

 and for various temperatures the formula 



v2=9 (t + x (To "" 0,03591 ^ +0, ° 531 ^) 



in which T is the expression within brackets of the first formula. 

 This latter value vanishes for X=28 centim. ; according to Thomson 

 the influence of surface-tension on this wave-length would still be 

 T =0*1 centim., that is about ^ per cent. For an infinitely small X 

 T = 0*070, which agrees sufficiently with the values ordinarily 

 assumed. It may be observed that with other methods the num- 

 ber of vibrations went down to 2- 67 and the wave-lengths rose to 

 21-915, while the T sank to 0*0130, and that nevertheless the above 

 two formulae for v 2 agreed very well. The actual state of the case 

 lies between Kolacek's formula and that of Thomson for X= 1*672 : 

 for instance, according to Thomson n = 13*83 ; according to 

 Kolacek, n = 12*33 ; experiment, n = 13*33. 



Phil Mag. S. 5. Vol. 31. No. 188. Jan. 1891. G 



