51(4) LRTS1 Subsequent studies have presented a wide range of fund allocation methods, including some of great sophistication. This paper reviews the literature on library fund allocation, then presents a systematic, statistically informed method of weighting and combining the variables in a fund allocation formula.The approach described here is most useful for identifying the relationships that underlie a set of previously established allocations-for revealing the formula that best matches the allocation levels set in previous years. It is therefore especially appropriate for institutions that already allocate funds based on historical precedent but without an explicit formula. Other libraries may find the method helpful as a means of evaluating and refining the formulas already in place. Specifically, the regression-based approach to library fund allocation can be used in at least three ways: to generate an allocation formula based on previous years' allocations (in those cases where funds have been allocated based on historical precedent without the use of a formula); to generate an allocation formula based on subjectively established allocations (in those cases where funds have not been allocated among departments); and to evaluate and refine the formulas already in use (in those cases where the current formulas are unsatisfactory or otherwise in need of modification).Rational, well-documented methods of fund allocation have several advantages over informal or ad hoc approaches. According to the Association of Research Libraries, allocation formulas and similar techniques promote transparency and the explicit recognition of underlying assumptions, encourage funding practices that are consistent with the library's goals and priorities, ensure that adequate fund monitoring mechanisms are in place, and help the library demonstrate to the university community how its funds are being spent.2 Fund allocation formulas also are likely to promote budgetary stability over time (i.e., to reduce the