OPTICAL PROPERTIES OF NUCLEIC ACIDS 543 



DNA solutions oriented by streaming and of gels oriented by stretching, 

 both with unpolarized radiation and with radiation polarized parallel and 

 perpendicular, respectively, to the orientation direction. Their theoretical 

 analysis indicates that the error in absorption measurements on such ori- 

 ented systems with unpolarized radiation will be dependent on the dichroic 

 ratio, and they suggest that the magnitude of the error must be investigated 

 experimentally for each particular case. Their results show small errors, 

 ca. 5%, for specimens with quite high dichroic ratios (up to 3) and high 

 transmission, but the error is very much larger for low transmissions. They 

 conclude, however, that the dichroism of the nucleic acid in cytological 

 material is so low, except for material such as sperm heads, that the effect 

 of molecular orientation on the absorption measured with unpolarized light 

 is unlikely to introduce the serious errors adduced by Commoner and 

 Lipkin. 



As noted above, Caspersson's early ultraviolet absorption measure- 

 ments^'* on an oriented film of DNA show good agreement between the 

 curve for unpolarized radiation and the average of the two curves for radi- 

 ation polarized parallel and perpendicular to the orientation axis. In this 

 example the average absorbance is ca. 0.6 (approx. 25 % transmission) and 

 the dichroic ratio about 1 .6; the experimental value for the peak absorbance 

 with unpolarized radiation is not more than 3 % lower than the calculated 

 figure. 



In a critical evaluation of quantitative cytochemical techniques Click, 

 Engstrom, and Malmstrom"® accept the conclusions of Thorell and Ruch, 

 and cite other examples of biological material in which orientation might 

 be expected but where the observed ultraviolet dichroism is low, so that 

 absorption measurements with unpolarized radiation should not be subject 

 to appreciable error. They also draw attention to the possibility of error due 

 to inhomogeneous distribution of absorbing material over the total area of 

 measurement.^" It can easily be shown that error due to such inhomogeneity 

 is most serious at high absorbance values. Since the proportional effect 

 of a constant fractional error in transmission on absorbance (and hence on 

 concentration) measurement also depends on the absorbance. Click et al. 

 emphasize the importance of working at fairly low absorbance levels. The 

 concept of an optimum absorbance range, approximately 0.2-0.7, for precise 

 spectrophotometry has long been accepted for measurements in solution, 

 where problems due to heterogeneity do not arise. "^■^'' 



In a discussion of errors in microspectrography Wilkins"" comments on 



"6 D. Glick, A. Engstrom, and B. G. Malmstrom, Science 114, 253 (1951). 

 '" R. N. Jones, J. Am. Chem. Soc. 74, 2681 (1952). 



"» G. F. Lothian, "Absorption Spectrophotometry," p. 52. Hilger and Watts, Lon- 

 don, 1949. 

 "9 N. T. Gridgeman, Anal. Chem. 24, 445 (1952). 

 i<° M. H. F. Wilkins, Discussions Faraday Soc. No. 9, 363 (1950). 



