Chapter 1 gives an overview and contains general information. Chapter 2 provides installation instructions and describes how to get started. It is recommended for those who are just beginning with eigenvalue computations and also for those who are unfamiliar with reverse communication. Chapter 3 gives a detailed description of how to use all of the capabilities of ARPACK. Those wishing to learn a little about the underlying numerical methods should turn to Chapter 4. This discussion provides a broad overview of the methods. It gives a reasonably detailed description of the Arnoldi process with implicit restarting and what to expect. It also attempts to provide some understanding of the spectral transformation. Numerous references are provided for those who desire a more detailed level of understanding. Chapter 5 discusses implementation and usage details within the main computational routines. Experienced users of large scale eigenvalue methods can probably turn directly to Appendix A and find the discussion of the driver routine that is appropriate for their problem. These drivers are intended to be used as templates that are easily modified for a particular application. Trace debugging and check-pointing are discussed in Appendix B. Finally, there are listings of the top level reverse communication interface routines XYaupd for reference. The source code for all of the computational routines is available with the distribution. Each of these is fully documented in the header and many users have found that documentation sufficient to get started with.