1081 
we have to solve the equations: 
and 
2(v7?—b)—6Y3be 2 —(Slat 4+ 27 b)etz + 482 ae — 0 
where 
This cubic equation has two positive roots, the first of which lies 
between 10-17 and 10—'8 and the second between 10-!® and 
9.10; these are of no use as they do not give a positive value for 
zy. The third root is negative viz: 
With this we find 
@ — 9.02 10—e2. NISA 1032, 
so that 
Pr Sid LOE: M = 2,0. 1016, 
If with these values we calculate the moment of momentum 
h 
mr’ M, we find for it: —. We are again led to the conclusion that 
oy, 
h 
it is impossible to obtain agreement with the value — sa introduced 
),a0 
by Bonr. 
§ 5. It would certainly have been better if the way in which 
we limit the freedom of motion of the electrons agreed more closely 
with the restriction which Bour has introduced for the electrons in 
the atoms, and which is one of his assumptions on which the 
deduction of BaLMER’s formula rests, viz. that for each electron 
. h 
mr D=. 
2m 
If in this way we find stable configurations, we have to solve 
the difficult problem to establish the equations of motion for a 
system, for which the equations of constraint contain velocities as 
well as coordinates. In dynamics we are taught how to form the 
equations of motion for those systems only in which the same con- 
nections exist between the velocities and between infinitesimal displa- 
cements (e.g. rolling without sliding) and this is not the case here. 
In order to investigate the stability we shall apply to this hydrogen- 
molecule the criterion, introduced by Bone instead of that of ordinary 
mechanies. This criterion is as follows: 
MOE 
