494 G. H. BEAVEN, E. R. HOLIDAY, AND E. A. JOHNSON 



such methods. A large proportion of spectroscopic work in the ultraviolet 

 region has however been carried out with a view to serving an author's own 

 particular analytical requirements, and both the scope of the work and the 

 manner of publication have sometimes reduced its general utility, resulting 

 in much unnecessary duplication. A critical survey of work in this field also 

 demonstrates forcibly that the use of compounds of a very high order of 

 purity does not of itself ensure the production of spectroscopic data of 

 comparable accuracy. The measurement of molecular extinction coefficients 

 to an absolute accuracy of better than 1%, or even 5% (Caster^), neces- 

 sitates careful attention to possible instrumental and operational errors 

 and to inter-instrumental discrepancies. These causes of error have re- 

 cently been discussed in some detaiP (Goldring et al}). 



The data given here are reproduced in accordance with the suggested 

 nomenclature of Brode^ and Hughes.^ The following symbols are used:- 



A = absorbance (optical density) = ecd = log y 



where e = molar absorptivity; 



c = molar concentration; 



d = internal cell length in centimeters; 



/o = intensity of incident radiation; 



7 = intensity of transmitted radiation. 

 Wavelengths (X) are given in millimicrons (mju). From the point of view of 

 theoretical interpretation of absorption spectra it would be preferable to 

 tabulate or plot frequencies rather than wavelengths, and if a wide spectral 

 band is to be covered such a plot also gives a more compact spectrum, 

 avoiding excessive spread at longer wavelengths. Since, however, the cali- 

 bration of all current ultraviolet instruments is given in terms of wave- 

 lengths, these are retained here. The manner of presentation of absorption 

 curves has in the past depended on the format favored by the journal in 

 question, which is, with a very few noteworthy exceptions, invariably 

 unsatisfactory from the point of view of the working spectroscopist. All 

 such curves should be given on an adequate scale, have coordinate grids 

 fine enough to permit accurate interpolation in both directions, and if 

 possible be printed in such a way that their scales have a metric basis to 

 permit direct measurement with an ordinary ruler.* It is rarely necessary 



* See pocket on inside back cover for scale drawings of Figs. 3 through 20. 



1 W. O. Caster, Anal. Chem. 23, 1229 (1951). 



^Photoelectric Spectrometry Group Bull., No. 3 (Oct. 1950). 



3 L. S. Goldring, R. C. Hawes, G. H. Hare, A. O. Beckman, and M. E. Stickney, 



Anal. Chem. 25, 869 (1953). 

 ^ W. R. Erode, /. Opt. Soc. Amer. 39, 1022 (1949). 

 6 H. K. Hughes, Anal. Chem. 24, 1349 (1952). 



