182 Dr. S. J. Lewis. Ultra-Violet Absorption Spectra and 



on varying the concentration of the ammonium sulphate is so small as not 

 appreciably to affect the specific rotations found. The effect under the 

 prevailing conditions of experiment is not likely to be greater than + 0*2 on 

 the specific rotation. 



Spectrophotometry of the Solutions. 



Strong solutions of the proteins were obtained by dissolving the purified 

 precipitates described above in water, and their concentrations were ascer- 

 tained by determining the total solids and the ammonium sulphate in the 

 solution, and taking the difference as protein. This method gave constant 

 and apparently satisfactory results. In every case the work was done in 

 duplicate, and sometimes in triplicate. 



The strong solution was polarised with the object of determining the 

 specific rotation, and suitably diluted with distilled water for use in the 

 spectrophotometer. The concentrations found to work best, that is, such as 

 to exhibit a well-developed band, have been found to be - 08 per cent, for the 

 albumin, 0'04 per cent, for the pseudo-globulin and for the eu-globulin. 



The strengths of the solutions were estimated approximately by means of 

 the polarimeter for immediate use, and corrected later when the chemical 

 figures became available. 



The dilution was filled into a 2-cm. observation tube fitted with quartz 

 ends, and a second tube was filled with a solution of ammonium sulphate of 

 approximately the same strength as was the protein solution with reference 

 to this salt : usually this was obtained by diluting saturated ammonium 

 sulphate solution 150 times. The latter tube was used as a blank in the one 

 path of the spectrophotometer, so that the observations made with the tube 

 of protein solution placed in the other path express the spectrum-absorbing 

 effect of the protein only. 



The process of spectrophotometry was conducted in the manner indicated 

 in the paper first cited, with all the refinements described in the paper dealing 

 with the new instrument. The series of photographs for each experiment 

 extended over three plates, making a series of fifty-four in all. 



The absorption curve is plotted with extinction coefficients as ordinates 

 and wave-lengths as abscissas. The extinction coefficient is calculated on a 

 1-cm. layer of a - l per cent, solution of the protein, which, according to 

 Beer's law, is the same as that on a 01-mm. layer of a 10 per cent, solution. 

 This corresponds with the protein concentration of serum as nearly as decimal 

 figures permit ; serum contains about 8 per cent of proteins. Hence the 

 curves approximate in their values to those already described with reference 

 to serum itself. In order to correlate the two sets of values, either the 



