Although salary compression has previously been identified in such professional schools as engineering, business, and computer science, there is now evidence of salary compression among Association of Research Libraries members. Using salary data from the "ARL Annual Salary Survey", this study analyzes average annual salaries from 1994-1995 through 2004-2005. It compares changes in salary ratios between entry-, mid-, and senior-level librarians for evidence of salary compression during this 10-year period for 16 of the 20 formally defined ARL position classifications. Evidence is found for salary compression among the position classifications, but, with certain exceptions, the phenomenon of compression has become less severe among ARL librarians over the 10 years surveyed. (Contains 7 illustrations and 13 notes.)