il86 



to oe at rest, bet-anse their inovenieiit may just as well be di rooted 

 towards the left .«ide as towards the right side (of. howevei- § 6). 

 Hence M moves steadily to and fro betweeji the molecules J/, and 

 J/, assumed to be fixed. We call u^ the velocity, which possesses 

 M in the point in the middle between M^ and M^ (indicated by 

 M in tig. 1 : the so-called dead or nenti-al point), either to the 

 lefthand side, oi' to the righthand side. 



With this velocity ^f covers the part MA^ (e.g. in /-*), but then 

 it enters the sphere of attraction (> of the molecule M^ (e.g. in Q); 

 i.e. we assume that the attractive force does not make itself appre- 

 ciably felt until the molecule has entered this sphere, and then 

 increases steadily rill .1/ touches the molecule M^ (distance of ihe 

 middle points = .y). Consequently the velocity u has also continu- 

 ally increased from A^ to a maximum value iis- Then repidsive 

 forces appear; the two molecules are a little compressed and M is 

 repelled. Between M^ ami M everything takes place in exactly the 

 same way, only in opposite order; and l)etween M and il/, and 

 back everything is again repeated. Hence we shall only have to consider 

 the fourth part, viz. the portion J7J/, of every movement to and fro. 



It is easily seen that three cases are possible, in the first place 

 the case represented in fig. 1, in which I ^ q, and the moleciile 

 M, therefore, passes over a longer or shorter path outside the in- 

 fluence of the atti-action of .lA,, and ahnays outside the sphere of 

 attraction of JV,. 



In the second place we have the transition case of tig. 2, vi/.. 



/ <; ^, but ]> ^ (^ -f~ •^)- Then M is continually within the sphere 

 of attraction .1/^ .4, of i17,, and besides during a period (from M 

 to A^, e.g. in P) also within that (J/, ^4 J of M^. As soon as xl/ has 

 passed the point A^, e.g. in Q, it will be outside the influence of il/,. 



/ 



Fig. 2. 

 In the third place 3/ can also be continually within the sphere 

 of attraction of J/, as soon as viz. M, on contact with M^ (distance 

 of the middle points =i s), is still just on the edge of Ihe sphere of 

 attraction of .17,; i.e. when A^M^=^1l — o has just the value s. 

 Then / is therefore =^ ,(?+•'')• Thus in the second (transition) case 

 ^]>V2(^+*'). while foi' the third case we have simply /-< 7, ((j4--''')- 

 Now we found before (see the cited papers) o to be about \,1 s. 



