The problem of the determination of trace quantities of impurities in metals has existed almost since the time of the first metal preparations of the cave men and exists today in the newest proposed alloys for high-temperature use or reactor construction. During this span of time only the definition of trace has changed. In early times, a trace might have been considered to be any quantity in the low percentage range; at present, we are most likely to define the term with reference to the field of endeavor in which we work. Current thinking in some metallurgical fields might define a trace to be 1000 ppm or less while in another area of operations it might well mean 1 ppm. In the field of semiconductors covered by James M. Morris, and Francis Pink, in this symposium, he might well refer to parts per billion as a trace impurity. In order to be able to limit the scope of this paper, the value of 100 ppm has arbitrarily been chosen as an upper limit for trace elements in metals.