900 
in which « changes in the above integrations from /—s’ to 0. And as 
s’ is always >-s (only for an infinitely large value of wu, could s 
be reached at the culmination of a collision), /—s’ is always < /—s. 
Kspecially at lower temperatures, at which s’ remains comparatively 
far from s, (/—s’): (/—s) can remain considerably smaller than 1 
even at the extreme value of z. (If eg. l=1,2s, s’=1,1s, this 
ratio becomes already '/,). We may therefore write in approximation 
for the logarithmic term: 
loo( 1 -— cs) Ze es et 
(/—s)’ (ls)? 2 (l—-s)‘ 
so that with 
(9 — s) (6—s) 
Pader: . . ° ° 65 ° Fy . (5) 
the following form is obtained for oi: 
ER NS (1 x ree ‘a 
Wy, = a tf) [ =f) (l—s)? an oe ee ( ) 
For the form under the sign of the root may therefore be written : 
| fas 2 
ws LE toer | |= we? Utp 
2 
Aa 
Ls 2f 
when again, as in the first paper, ee is put, and further 
Us, m 
. . x 
y is substituted for ja Then: 
Is’ l—s' 
Ls l—s 
ls (° dy 
Ae (te) fra. dy. 
Jay's 
For the form under the sign of the root (1—v2,y?) (1 + w,7?) may 
be written in this, when 
w, — w, =p (le) ; vw, vw, ='/, PO, 
. . . w 
so that this form with yw, =z passes into (1—z’) ( +— Z|, 
w, 
which becomes with z=cosw: 
3 Ws Î als ae os Ws 
sin? W (1 + — (1 — sin’ w)=' sin wl = — sin ‘w ) 
w, ww, 
0, 
. dz 
For dy we have further —— = — 
1 
Vw, Vw, 
sin w dw, so that the 
above integrals pass into 
l—s BA 
== — 
u Vw, w ite, ln ‘1—F sin? wy a 
