Amos Ron's personal comments: the list is obviously tilted towards applied math, as any Citation count should. Newer journals like JAMS are not to be found. Keep in mind that the low rank of J. Approximation Theory reflects its status before Nevai and Pinkus took over. It is heartening to see the high number of citations of Constructive Approximation, despite its specialty character and newer age.

Anonymous note: Note that the relative ranking of a journal $X$ in pure mathematics is obtained by the following formula: let $S_X$ be the set of all the papers published in journal $X$ in the period 1983-1992 and $c_X$ the number of citations of the papers in $S_X$ cited by mathematical publications in 1993; $f_X$ is defined to be $c_X/|S_X|$ and the journals are ranked according to their $f_X$ values. For journals in applied mathematics, mathematical statistics, and applied statistics, final $f_X$ values are computed by taking some representative journal in each category to be equivalent to some journal in pure mathematics. Since the $c_X$ values provided by SCI and the $|S_X|$ values searched from Math Sci System may not be reliable, the ranking for pure mathematics is thus not necessarily accurate. Since the method for the relative ranking of journals in applied mathematics, statistics and applied statistics among all the journals in mathematical sciences considered here is based on the rather arbitrary choice of a representative in each class, the ranking for these journals not in pure mathematics is less objective.