Manchester Memoirs, Vol. Iviii. (19 14), No. 15. 7 



Source of Light. The spectra of most, if not all, 

 sources of light contain weak regions where both the 

 maximum intensity of individual lines and the average 

 intensity of the region are less than in the surrounding 

 regions. There is thus a tendency to the formation of a 

 spurious band at this point, and where a genuine band 

 occurs at this wave length the effect will be liable to 

 exaggeration. 



Exposure. Under average conditions an exposure of 

 15 to 20 seconds is sufficient to produce a fairly dense 

 and readable plate. Changes in the time of exposure 

 do, however, greatly affect the position of the curve with 

 reference to the concentration. Bands tend to become 

 narrower with longer exposure. 



Thus, since no convention exists as to a light source 

 of standard intensity, it is only by chance that the curves 

 for a given substance obtained by two independent 

 workers will be perfectly superposable. As a rule, for the 

 purposes for which the method has mostly been used, 

 this is not serious, but it is a fatal objection to methods 

 of analysis based on absorption spectrum measurements. 



Extra Absorption. Three sources of error are summed 

 up under this term, absorption by the solvent, scattering 

 of ultraviolet light by fine suspended impurity, and 

 absorption in the region \ = 3,000 — 2,000 by the materials 

 of the plate itself The second may easily be remedied by 

 good filtration, but is often overlooked, the first and the 

 third are serious. Hartley in his paper quoted above 

 gives a list of solvents transparent to ultraviolet light 

 down to A = 2,000. This list appears to have been 

 accepted without question by chemists. But it is easy to 

 show in every case that weakening of the transmitted 

 spectrum occurs as A = 2,000 is approached. To take the 

 case of alcohol ; an ordinary specimen of Kahlbaum's 



