Another Look at Conditionally Gaussian Markov Random Fields

                          Michael Lavine

             Institute of Statistics and Decision Sciences
                      Duke University


  This paper shows that a Markov random field prior on a lattice is
  identical to the posterior from a particular dynamic linear model
  updated with a particular data set Y.  This opens the
  possibility of using dynamic linear model ideas, notably a
  computational algorithm, in the Markov random field setting.