DISCUSSION AD CONCLUSIONS
This work shows that soils collected over fairly short distances have a variation in trace elemental composition. It was shown that neutron activation analysis and atomic absorption are valuable adjuncts to the conventional methods of forensic soil examination. The many elements that can be determined by these techniques provide the points of identification necessary for the comparison of questioned and known soils.
It should be noted that in this control study, every effort was made to sample areas where small differences in trace element concentration would be expected. This was accomplished by collecting samples from relatively small flat areas which appeared to be uniform. Three times the analytical precision was used to establish the concentration ranges observed for each element, and only those ranges occurring most frequently for the elements were used to calculate the probability of randomly selecting two samples containing 15 elements in the same concentrations. Even with liberal probability factors, the variations in the soils connected with this study show that specimens from different locations can be distinguished.
These calculated probability factors are not applicable to all soil comparisons because it is necessary to establish the concentration ranges for each element from a particular collection area. Their use in this work is only to illustrate the magnitude of the differences in soil composition when a large number of trace elements in specific concentration ranges are employed as points of identification. In an actual case, where a limited number of soil samples are collected, an estimate of the probability of finding two soils from different locations with the same trace elemental composition may be expressed by a simplified relationship: