Version 1.3.3 (6/30/15):

* fixed bug in propose.coeff.gp in line 30, which is now:
  object$coeff[1,1]=Re(object$coeff[1,1])

* fixed bugs in propose.coeff.gp, Gibbs.sample.coeff.R, and Hastings.coeff.gp.R
  in which I was using object$variance.param*object$variances even though 
  object$variance.param was already included in the calculation of 
  object$variances.

Thanks to William Leeds for pointing these out.

* changed the image.plot and related functions copied from fields to be 
  named with _ to avoid R thinking these are S3 methods 
