of Rubidium by the Spectroscope. 



395 



When the volume of the test-solution reaches 400 cm 3 the 

 indications are, on the whole, that the solution is still stronger 

 than an equal volume of the standard ; and when the volume 

 amounts to 420 cm 3 it is safe to conclude from the indications 

 that the test-solution is the weaker. If the mean of the num- 

 bers representing these two volumes be taken as probably indi- 

 cative of the volume at which the test and standard lines are at 

 an equality we have the estimated amount of rubidium in the 

 test-solution given by multiplying the volume in cubic centi- 

 meters by the number of coil-fulls in 1 cm 3 and the product 

 by the amount of rubidium contained in a coil full of the 

 standard solution— 410x50x0-0005 = 10-25 mg. The amount 

 of rubidium actually taken in the test-solution was 10 mg., and 

 the error of the mean value assumed is 2 5 per cent + ? 

 between extremes corresponding to an error of and 5 per 

 cent -f , respectively. 







Experiment II. 





Standard. 



Rubidium in 



a coil-full (a 1 ,,- cm 3 .) 





Test (known to 



contain 10 mg. Rb.). 



Volume in cm 3 . 



Line of test 



compared with 



standard. 



0-0005 mg. 



(C u 



u a 





340 

 370 

 370 



Brighter. 

 j Equally bright. 

 \ Brighter. 



a u 





390 

 390 



j Weaker. 

 | Weaker. 



Found — 



9, 



390 



X50X0-0005 : 



= 9 - 5 mg. 



Taken 







10-0 " 



Error 







0-5 " = 5 per cent. 







Experiment III. 



Standard. 



Rubidium in 



a coil-full (g^- cm 3 ). 





Test (known to 



contain 10 mg. Rb.). 



Volume in cm 3 . 



Line of test 



compared with 



standard. 



0-0005 mg. 

 u u 



a a 



a a 



a a 



a a 





300 

 360 

 380 

 380 

 390 

 400 

 410 



Brighter. 



Equally bright. 



Brighter. 



Brighter. 



Brighter. 



Weaker. 



Weaker. 



.bound - 



+ 

 2 



400 



X50X0-0005 



= 9-875 



Taken .... 







io- 



Error 







0-125 = 1-25 per cent 



