552 On a Kinetic Theory of Solids. 



s, then A the parameter of molecular force is proportional to 

 s/mtp*. But if (j>(r) = 3Am 2 /V 4 , then 



%r<j>(r) = 12Amirp log E/a, 



where log E/a is nearly the same for all bodies not excessively 

 small or large; hence 



2r<£(r)/6e 3 = 2A/) 2 7r log U/a = 2T)/3eXe-E) = 2Jcmp/21mb, 



or A is proportional to cm/mbp, or, as cm is constant, to 1/mbp. 

 Thus s/m* p^ is proportional to I'fmb, or bsm$/p* is constant 

 It is to be noticed that while 2r$(V)/6e 3 or 2Jcmp/21mb is 

 independent of the molecular mass, the surface-tension 

 a; Am^ ^1- | s no t ? so that if the molecular weights of the metals 

 are not all the same multiple of the atomic weights, excep- 

 tional metals will give exceptional values of bsmj/p%. 



The following is a list of the values of 1000 bsm?/p%, s being 

 given in milligrammes per millimetre. Where Quincke 

 found two values of s by different methods, both are given. 

 In the case of lead the surface-tension was determined by 

 myself. 



Pt. Pd. Au. Au. Ag. Ag. Sn. Pb. 



s 169 136 100 131 43 80 60 34 



1000 fom*/p*... 7-3 7-4 7*4 9'7 4-0 7-5 8*7 6*9 



Zn. 



Od. 



Na. 



K. Hg. 



Fe. 



Fe. 



83 



71 



26 



37 59 



102 



97 



10*8 



13-6 



14-4 



26 28 



4-7 



4-4 



S ss 



1000 bsm%/p$... 11-4 



The value of the surface-tension of iron is uncertain, as is 

 the case also with sodium and potassium ; for the value of b 

 for Hg, Na, and K, I used the relation &Tm* = "044. The 

 mean value of 100Qbsm$/p% for Pt, Pd, Au, Ag, Sn, and 

 Pb is 7'4, for Zn and Cd it is 12'3, which is once and a 

 half as large or about 2^ as large, for Na it is twice as large, 

 and for Hg and K four times. These relations and the 

 whole subject of surface-tension in melted solids require further 

 investigation. I have touched on the matter here to show 

 that in the kinetic theory unfolded, the law of molecular 

 force has not been lost sight of. 



There are many phenomena of solids that invite investiga- 

 tion by the light of a kinetic theory, but it has seemed better 

 to me to confine the present paper to the broad fundamental 

 ones. One of the results of the inquiry is that the molecules 

 of solids so nearly fill up their domains that the phenomena 



