THE STRUCTUKE OE TILE NUCLEUS. 



45 



being more complete will be made the })asis of coinimtatiou. It remains extremely 

 difficult to observe the faint axial colors of lii<rher orders, 2, for which reason 

 differences of color for large z are inevitable. Moieover, as all eii-ors on exhaustion 

 are cumulative, the general coincidence of the two sei'ies is rather suiprising. 



As in § 13, let iV be the nuniher of nuclei after the 2th exhaustion. Assum- 

 ing that both the normal time losses and those removed by exhaustion follow the 

 same exponential law, put 



where h is the coefficient of the time loss and y the ratio of the geometric progres- 

 sion of exhaustions, computed either isothermally N-^ or adiabatically N.^. In the 

 first case, y =p/Pq = .888 ; in the second, y = (p/p^y/y— .919. 



TABLE 19.— AXIAL AND CORONAL COLORS (Central Field) SEEN IN THE DOUBLE 



DRUM. PUNK NUCLEI. EXHAUSTION, 76-67.5 cm. 



t=3-5 rain., l> = .io. y= (///„) = .888. 



logiV=s(i+<5/) logj.. 



y 



i/v 



.919. 



24. Loss of nuclei in the lapse of tdme. — To complete the computation, it is, as 

 above, necessaiy to make special experiments on the time losses. These are given 

 in tables 20 and 21, for exhaustions between 76 and 68 cm. 



Thus, after long standinof, the effect of identical successive exhaustions is more 

 rapid than in the preceding table. Moreover, after two hours certain qualita- 

 tive variations, apart from mere loss of number, seem to occur. There is again a 



