Artificial isochrons

Abstract

The rubidium-strontium isochron procedure has proven to be a powerful tool in many geochemical studies. An isochron is said to exist if a plot of ( 87Sr/ 86Sr) versus ( 87Rb/ 86Sr) produces a linear trend with a large correlation coefficient. However, the variables selected to portray the isochron have the same denominator ( 86Sr) and Pearson long ago noted that a large correlation could be induced when such ratios are formed from uncorrelated numerators and denominators. A set of numerical experiments are described that illustrate the common denominator effects as applied specifically to rubidium-strontium systematics. For at least one previously published RbSr isochron it can be shown that the common denominator effect is capable of producing a correlation coefficient that is very nearly 1.000. However, it is also shown that, for the data sets analyzed, the common denominator effect can not produce a geologically meaningful isochron. The numerical approach to assessing the common denominator effect can be applied only to those sets of isotopic analyses in which 87Rb and 86Sr have been determined by isotope dilution techniques. For the many data sets in which only the ratios ( 87Sr/ 86Sr) and ( 87Rb/ 86Sr) have been determined the common denominator effect can neither be assessed nor dismissed as trivial.