407 







(6) 



Taking A, = 4500, ;.,= 6870 we liud from (6) —'=i 1,572. from 



A,. 



A, " L, 



(5) -—^==1,608, whereas the experiment (Table p. 403) gives — ^=1,616. 



For' ;ii = 4580, A, = 6820 the ratios become respectively 1,542, 

 1,575, 1,581. 



So there is only a difference of 0,5 resp. 0,4 7,, between the 

 formnla of I^okkntz ami the experiments, but a difference of 2,2 

 resp. 2,0 "/„ between these and tiie formnla of Fkksnkl. ') 



Even if wo had not sncceedcd in giving to /, 't'„,„, and the co- 

 efficient 0.84 very probable values, e\fn then the result of our ex- 

 periments had been very favourable to (-(luation (5). 



Further we must mention, ilial the light beam was limited b}' 

 rings of tin-foil to a width of II ni.m.. whereas the glass plates allowed 

 a beam of IH m.m. diameter to pass along the axis, the horizontal 

 tubes through which the water flows being of an inner diameter of 

 40 m.m. By this precaution the optically efleclive change of the velo- 

 city over the section of the tube is diminished and this is also the 

 case with the broadening of the interference fringes caused by the 

 curving ot the wavefronts by inequality of the velocities in them. 



Sometimes (not always) there is a snnill change in the distance 



between the interference fringes after reversing the dii-ection of the 



water current. It is easily proved, that, neglecting 



?' quantities of the second order, wo get a right result 



i by dixiding the mean value of the distance between 



I ^ J i I <f the interference fringes before and after change of 



' the current in the shift of tiie fringes. 



Let ff, and *', be the distances between the fringes 

 i in the two cases and a^ and a, the shifts of them 



j from the original position 00'. PVom the measure- 



O . , ''l + «3 



Fig. Ü ments we hnd or rather the double of this. 



'^i + rf, 



Wc want to know — = . The difference between the first and the 

 second expression gives the error we make. Let us put 



') Our conclusion is confirmed by a recent, more accurate series of observiUions. 

 [Note tu the translation]. 



