List of typos in first and second printings of ``Hierarchical Modeling
and Analysis for Spatial Data,'' by Banerjee, Carlin, and Gelfand:

  (In all cases below, ``s/b'' = ``should be'')


p.8:  Equation (1.3):  \phi_i inside summation s/b \phi_j
  Also just below (1.4):  A = \rho W s/b A=\rho Diag(1/w_{i+}) W

p.9, line 17-:  ``regularly space pattern'' s/b ``regularly spaced pattern''

p.11:  Add to figure caption:  ``(figure courtesy of the U.S.
  Geological Survey)''

p.13:  first display formula:  f(\lambda, \phi) = R \lambda

p.15:  Add to figure caption:  ``(figure courtesy of the U.S.
  Geological Survey)''

p.16:  Add to figure caption:  ``(figure courtesy of the U.S.
  Geological Survey)''

p.18:  equation (1.10):  last term s/b cos(\lambda_1 - \lambda_2)

p.22:  line 4:  two s's s/b boldface

p.24:  figure caption:  ``a0'' s/b \tau^2, and ``a1'' s/b \sigma^2

p.25:  line 4-:  ...effective range *is* often used... 

p.26:  fraction in expression for rational quadratic s/b
  \frac{\sis t^2}{(\phi + t^2)}

p.27:  Table 2.1:  fraction in expression for rational quadratic s/b
  \frac{t^2}{(\phi + t^2)}

p.28:  Table 2.2:  same fix as p.26

p.33, line 8+:  exponent -2 s/b -r

p.34, line 13+:  delete reference to Cauchy correlation function
  (not included in Tables 2.1-2.2)

p.37, line 6-:  "in the development" s/b "on the development"

p.38:  display formula halfway down the page:  two s's s/b boldface
  Also, line 6-:  "as $c \rightarrow 0$, this says that..."

p.46, line 6-:  ``polar cordinates'' s/b ``polar coordinates''

p.47, line 3-:  ``...using empirical semivariogram values in the row of 
  the ESC plot for which $h_y \approx 0$ provides an alternative...''

p.51, line 1+:  \beta should be bold

pp.61-65:  The R language no longer permits use of the underscore
  for assignment, so these should now be arrows (``<-'')

p.61:  Change the ``interp.new'' command to simply ``interp''
  Also, delete the ``extrap=T'' option from this command

p.62:  Initialize the geoR package by typing ``library(geoR)''

  Also, the second geoR command should be:
    scallops.geo <- as.geodata(obj,coords.col=1:2,data.col=3)

  Also, in the third and fifth geoR commands, change ``variogram'' 
    to its replacement, ``variog'' 

p.65:  The second geoR command command should be:
    out <- scallops.bayes1$posterior

p.66, Problem 3:  add square brackets around the 1-cos(\delta x)
  Also, Problem 4:  "...and let W(t) be a *mean zero* process on..."

p.73, expression (3.3):  denominator should be multiplied by 2

p.76, line 10-:  ``i.e., $E(Y_1)$ and $E(Y_2)$ are linearly related.  But...''

p.83, expression (3.20):  w_{ij}  s/b  b_{ij}

  Also, line 12-:  bold \beta s/b bold \gamma

  Also, expression (3.21):  \gamma s/b bold \gamma

p.84, line 16+:  *the allocation model in* Green and Richardson...

p.85, line 19-:  Y = X \beta  s/b  Y - X \beta

p.93, second R command:
  cutoff.sids <- c(0.0,2.0,3.0,3.5,7.0)    s/b
  breaks.sids <- c(-0.001,2.0,3.0,3.5,7.0)

pp.94-95:  ``cutoff.sids''  s/b  ``breaks.sids''

p.96, lines 1+ and 2+:  ``prior''  s/b  ``specification''

p.96, lines 3+ and 6+:  ``nonsingular''  s/b  ``positive definite''

p.96, line 7+:  \rho > 0  s/b  0 < \rho < 1

p.112, line 4+:  by  s/b  bold y

p.112, line 11+:  ``regulatory''  s/b  ``regularity''

p.132, 2nd para:  p(\bW | \sigma^2, \phi) s/b  p(\bW | \btheta, \by),
            and   p(\sigma^2, \phi | \by) s/b  p(\btheta | \by)

p.134, line 2-:  Y0[j]  s/b  mu0[j]

p.143, line 3-:  ``to minor axis (say, from a rose diagram), and a
  guess...'' 

p.147, line 5-:  ``this is not *necessarily* sensible, since...''

p.156, expression (5.38):  y-s  s/b  y-x

p.157, line 8-:  ``a continuous process *that is non-Gaussian*.  But...''

p.181, line 4+:  the s in Y(s) should be bold

p.217, line 1-:  the s's in C(s,s') should be bold

p.218, line 17-18-:  Gelfand, Schmidt, *Banerjee*, and Sirmans (2004)
  (see also reference list, pp.430-431)

p.219, line 7+:  The T should be italic

p.221, line 10-:  The Y should be italic

p.223, line 11-:  ``Wishart'' s/b ``inverse Wishart''

p.227:  IG(2, 0.024) s/b IG(2, 1/0.024)

p.228, Table 7.1:  ``\rho (correlation coefficient)'' s/b
  ``T_{12}/\sqrt{T_{11} T_{22}} (nonspatial correlation coefficient)''

p.244, line 3+:  ``In this expression, Z(\cdot) is a mean 0 stationary 
  process with correlation function \rho(\cdot), and k_{l} is ...''

p.257, beginning of last line of 2nd para:  
  \widetilde{Y}_{rows} s/b \widetilde{Y}_{cols} 

Also, same line:  \widetilde{Y}_{rows}^T \widetilde{Y}_{rows}^T
  s/b \widetilde{Y}_{cols}^T \widetilde{Y}_{cols}

p.260, last line:  italicize ``process'' instead of ``dummy''

p.261, 2nd line of Sec 8.1.3:  t \epsilon {1, 2, ..., T} 
  s/b   t \in {1, 2, ..., T} 

p.263, 2nd para, 3rd line:  ``observed vector of log selling prices''
  s/b ``observed data vector''

p.263, (8.17):  \times f(variables)  s/b  \times d F(variables)

p.276, second display formula:  \beta_t  s/b  bold \eta_t

pp.277-280:  All locations s should be bold

p.278, halfway down:  ldots s/b \ldots (i.e., ...)

  Also, line 13-:  add, ``where $[R_{l}(\phi_l)]_{ij} =
  \rho_{l}(\bs_i-\bs_j ; \phi_l)$ is the correlation matrix for 
  $v_{l}(\cdot,t)$.  We then write...''

p.351, line 21+:  ( D_{10} (\bs_i) , D_{01} (\bs_i) )  
      s/b         ( D_{10} Y(\bs_i) , D_{01} Y(\bs_i) )  

p.367, lines 7- and 11-:  sample size I s/b n
  Also, line 1-:  X(\bs)  s/b  X_{s}

p.371, line 3-:  1/I  s/b  1/n

p.416, line 5-:  \phi1 and \phi2  s/b  \phi_1 and \phi_2

p.452:  Index entry for ``spatially varying coefficients'' should also
  reference p.321