This document describes a library of Splus functions used for nonlinear Bayesian regression in general and IR estimation in particular. This library has been developed to solve a general class of problems described by the nonlinear regression model: Y = F (beta,data)+ E where Y represents a vector of measurements, and F(beta,data) represents a Splus function that has been constructed to describe the measurements. The function F(beta,data) depends upon beta, a vector of parameters to be estimated, while data$ is an Splus object containing any other information needed by the model. The errors, E, are assumed to be independent, normal, unbiased and to have known standard deviations of stdev(E) = sd.E. The components in beta are split into two groups; estimation parameters and nuisance parameters. The Bayesian prior on the estimation parameters will generally be non-informative, while the prior on the nuisance parameters will be constructed to reflect the information we have about them. We hope an extended beta distribution is general enough to adequately represent the information we have on them. While we expect these functions to be improved and revised, this library is mature enough to be used without major modification.