( 536 ) 



TAHLE 111. 



— U.ll = 10.13 a\ 4- G.7G </., -f 2.49 a'^ 



— 0.04 = 8.97 (i\ + 3.07 a'.^ — 2.74 a\ 



— 0.02 — 6.51 a\ — 3.07 a'., — 8.83 a\ 



— 0.01 — 3.26 a\ — 8.40 a'.^ — 7.51 a', 

 —'0.01 = 0.00 a\ — 10.38 a'., + 0.00 a\ 

 _ 0.02 = — 3.19 a\ - 8.41 a', f 7.52 a', 

 _|_ 0.04 = — 6.47 rt\ — 3 05 a\ + 8.81 a, 

 -I- 0.09 = — 8.88 a\ + 3.04 a^ -f 2.73 </, 

 _^ 0.1 P-— — 10.42 a\ + 8.83 «', — 5.42 a', 



TAIU.K \\\}> 



+ 5.14//, 

 + 0.00 //. 



- 6.24 h\ 

 + 0.00 h\^ 

 -y 6.24 b\ 

 -f 0.00 //, 



— 6.25/^', 

 + 0.00 //, 

 + 6.26 b\ 



Hj means of the metliud of least s(iuur('S '), we liiid the normal 

 equations eombined in tahle 1\. 



TABLE IV. 



475.8 a\ — 17.92 a\ — 130.65 ./, + 3.8284 =: 



— 17.92 a\ + 401.4 a', — 28.31 a\ — 0.6606 = 



— 130.65 a\ — 28.31 a^ + 319.0 «', -f 0.0882 = 



TABLE IVA. 



— 13.31 h\ 



— 60.33 //, 



— 21.02//. 



— 13.31 a' 



00.33 a', — 21.02 </, + 182.6 //, + 0.0331 







These equations yield 



— 0.00908 , 

 0.000964, 



t 



(4) 



a'3 = — 0.00391 . 



By means of the e([uation (3) we can now cahMdate the values of 

 L lor the different values of M in order to judge whether they 

 agree sufiieiently with the values given by the experiment. Then in 

 the second number we may assign to A the values of table I, and 

 so use the coeflicients given in table III, as these terms have little 

 intluence on the result. It now apjjeared that it was advantageous 



1) Although each of the equations (3) contains 2 quantities dedueed from obser- 

 vation, I have not applied here the method described in Supplement N" 4 to tlie 

 Clonimunications from tlie Phys. Lab. of Leiden, These Proc. V Sept. 27. 1902 p. 

 !23r» on the reduction of equations of observations containing more than one measured 

 quantity, because 3/ in comparison with ,-, may lie supposed to be accurately known. 



