310 
sented by the dotted lines). Indeed, for the sake of simplicity we 
have so far always neglected the influence of z. 
We still point out that the magnitude and the sign of the action 
exercised are always represented in the figures by the inclination 
of the tangents to the curve. 
2. 0 =180°. Then the electron passes the nucleus (fictitiously) 
just from the left to the right, when M goes from O to P, and (8) 
becomes : 
They t 
lily = ee Na ORO) ; PT Rn A 
Now the velocity w always remains below wu, (see Fig. 4). The 
case of “collision” has been drawn at two successively possible 
places, viz. at A and A’. 
What distinguishes this case from the preceding one, is the possi- 
bility that wv becomes =O before the “collision”, and the molecule 
Oz /Po° 
=m ee e2@ © @ «= © @ ew ew wee = 
X=0 ~& o a 0 = a 
Fig.” 4. 
accordingly already “returns” before P has been reached. This will 
evidently take place as soon as wu, is so small that M lies low 
enough for the curve to intersect the t-axis (u — 0) (Fig. 4a). This 
takes place eg. in B. Transformed spatially this means that the 
molecules will move round the position of equilibrium O in closed 
orbits, as soon as we get below the point where the curve touches 
the ¢-axis for the first time (melting point)'). In this case the 
Fig. 4a. 
') I may be allowed to anticipate on what follows, and state here that the 
melting-point calculated in this way for H — if it were realisable — willlie 
at 36°,4 abs. As this melting-point must lie higher than that of H, (because 
the molecular attraction a that plays a part in it, is greater for H than for 
H,), this result is not impossible. (melting-point H, lies at 14° abs.). 
