with the strata and that the sequence of many such patterns provides a unique 

 characterization in stratigraphy. 



The thermoluminescence of Pennsylvanian limestones in Iowa, Missouri, 

 and Kansas has been studied by Bergstrom (1953). He concluded that a glow- 

 curve pattern may persist in a single thin bed and that successive layers in a 

 vertical section can give cyclical variations. The chemical composition of the 

 seas at the time of deposition appeared to be more important than the radioactiv- 

 ity as determined by the alpha counts. There was less variation in glow curves 

 among samples taken deep in a basin than among samples taken near shore 

 deposits where the chemical composition was subject to greater differences. 

 Whenever there were lithological changes in the limestones, there were likely to 

 be changes also in the glow-curve patterns. The quantity of acid-insoluble resi- 

 dues was measured, but is was difficult to establish firm correlations in thick 

 widely separated beds. 



Parks (1953) determined the thermoluminescence glow curves of surface 

 samples and subsurface cuttings of the Chester series of Illinois, Indiana, and 

 Kentucky. He used the heights of the high-temperature peaks and the ratio of 

 the highest peak to the next lower peak as a means of correlation. There was 

 considerable variation, but the sequence of glow-curve types in a vertical section 

 was fairly consistent and could be used for approximate correlation. 



Pitrat (1956) studied several limestones including the Madison limestones 

 in Utah. He stressed the importance of taking unweathered samples which were 

 as free as possible from secondary calcite and silica deposits. He kept the samples 

 on dry ice after irradiation with gamma rays and in most samples observed three 

 peaks. In a characteristic sample, the lowest peak at 75C had a relative height of 

 10, the 200C peak a height of 6.7, and the 300C peak a height of 3.2. In com- 

 paring the glow curves of the different samples, he tried relative peak heights, 

 planimetered areas under the curves and ratios of the peak heights. The best 

 results were obtained by comparing for the different samples the double ratio 



peak height 1 . peak height 2 



peak height 2 peak height 3 



where numbers 1, 2, and 3 refer to the peaks at 75C, 200C, and 300C respectively. 

 He also used a system of moving averages in which for a series of lateral samples 

 he would take the average of the first, second, and third samples, then the aver- 

 age of the second, third, and fourth samples, and so on. He concluded that the 

 variations are so great that detailed close sampling is not worthwhile, but he 

 achieved some success in general correlations using these averages. 



199 



