John Mattingly will present an experimental application of Bayesian inference to the problem of wide area urban source search.  This application seeks to estimate the a-priori unknown location of a radiation source from a collection of detector measurements in an urban environment. Prior work has investigated the application of trilateration, maximum likelihood estimation, and Bayesian inference to similar source localization problems; however, the prior work does not address the complication of estimating source location when highly attenuating objects (i.e., buildings) occlude the detectors’ view of the source. NCSU has coupled a simplified model of gamma transport through a heterogeneous medium with a Markov Chain Monte Carlo sampling procedure to estimate the posterior distribution of probable source locations given a collection of detector measurements, and we have evaluated the accuracy and precision of this method for source localization in an experimental application designed to mimic the conditions encountered in wide area urban source search.