This paper addresses multi-source information integra- tion in stochastic environments where information from sources consists of probabilistic domain models (repre- sented as joint distributions or Bayesian networks) learned from data. We extend the batch algorithm proposed by Maynard-Reid II and Chajewska [9] to accommodate incre- mental integration so as to support `anytime' querying. Ex- perimental results verify that our algorithms compare well with the batch algorithm in accuracy and efficiency. Ex- tensions for integrating joint distributions are independent of the order in which sources arrive, but the Bayesian net- work integration extension is only approximately so. This is due to bias introduced by the algorithm's use of heuris- tic optimization, and an `inertial' effect that makes this bias difficult to undo over time.