`Mixed-Effects Models in S and S-PLUS' by J. C. Pinheiro and D. M. Bates
==============================================

Errata and comments for the first printing (May 2000).

p. iv	Douglas Bates's address should be "Department of Statistics",
	not "Department of Computer Sciences and Statistics".

p. 15   The interpretation of the coefficients (betas) in the Helmert
        parameterization is incorrect. It should be:
	- \beta_2 -- Half the difference between T1 and T2.
	- \beta_3 -- One third of the difference between T3 and the
        average of T1 and T2
	- \beta_4 -- One fourth of the difference between T4 and the
        average of T1, T2, and T3.

p. 20   In the first paragraph following the output of
        intervals(fm1Stool), \sigma_b can vary by a factor of 3 (not 5). 

p. 24   In the last paragraph, the more specific model is fm1Machine.

p. 27   First sentence after first S output, only \sigma, and not
        \sigma_2, seems to vary over twelve orders of magnitude.

p. 27   Second S output, 0.79621 should be replaced by 0.99958.

p. 35   Last sentence of first paragraph, missing period at the end.

p. 58	In equation (2.1), the variance-covariance of b_i should be
	\Psi,  not \Sigma.

p. 60	Line -3, the variance-covariance of the b_i, b_ij, and b_ijk
	should be \Psi_1, \Psi_2, and \Psi_3, not \Sigma_1, \Sigma_2,
	and \Sigma_3.

p. 63   In Equation (2.5), the first exponent should be divided by 2.

p. 65   Last sentence of first paragraph, missing "to" in "If possible
        we want to take advantage ..."

p. 68	Line -2 of the first paragraph: ... is the rotated response
	vector. (not residual vector)

p. 69   In Equation (2.15), the first denominator should be to
        (2\pi\sigma^2)^{q/2} instead of \sqrt{2\pi\sigma^2}.

p. 69   There is a missing `i' as a subscript of the last Z under the
	squareroot in the last display.

p. 70   First displayed formula, second line. The exponent "-N/2"
	in the denominator should be "N/2"

p. 72   First paragraph after numbered list, missing period at the end
        of the sentence.

p. 84   Last sentence of first paragraph, the REML version of BIC also
        substitutes log(N) by log(N-p) in (2.28). 

p. 150  Line -4, should be "... while F03 is above the girls' average."

p. 160  The second matrix in the Kronecker product defining the X_i
        matrix should be

	             | 1  N_1 |
		     | 1  N_2 |
		     | 1  N_3 |
		     | 1  N_4 |

p. 169  Line -10, there is an extra "that" -- should be "... confirms
        the significance of ..."

p. 178  Line -2 of first paragraph, missing period at the end of the
        sentence starting with "Overall, the standardized residuals...".

p. 186  S output for intervals(fm2Wafer), the confidence intervals for
        the standard deviations were calculated with a version of NLME
        that contained a bug. The correct intervals are

	. . .
            Random Effects:                             
	     Level: Wafer                               
	                       lower     est.    upper  
	    sd((Intercept)) 0.080154 0.128884 0.207240  
	        sd(voltage) 0.213234 0.348651 0.570065  
	   sd(I(voltage^2)) 0.029014 0.049074 0.083002  
	     Level: Site                                
	                      lower     est.    upper   
	    sd((Intercept)) 0.02195 0.039675 0.071712   
	        sd(voltage) 0.19087 0.234373 0.287790   
	   sd(I(voltage^2)) 0.03829 0.047541 0.059028   
	                                                
	    Within-group standard error:                
	        lower     est.   upper                  
	    0.0092741 0.011325 0.01383                  

p. 197  Line 6, should be "... introduced to model ..." instead of
        "...introduced the model...".

p. 206  Line -6, missing "be" -- should be "... can no longer be
        assumed to be independent."

p. 231  In the second equation, h_{nugg}(s, c_0, \rho) = 1, when s = 0.

p. 255  Caption of Figure 5.17, "plots" is repeated twice.

p. 274  Line 2 of third paragraph, the value of \phi_3 should be 0.2
        instead of 1.2.

p. 277  Section 6.2, "indomethicin" should be replaced by
        "indomethacin" throughout this section. 

p. 296  Line -5, should be (6.8) instead of 6.8).

p. 343  First sentence of third paragraph, should be "... by building
        a function for the logistic model".

p. 373  First line: ... asymptotic uptake rate. (not "update" rate)

p. 400  S output for intervals(fm5OvarlmeML), the confidence intervals for
        the standard deviations were calculated with a version of NLME
        that contained a bug. The correct intervals are
        . . .
          Random Effects:
           Level: Mare 
                                   lower      est.    upper 
               sd((Intercept)) 1.2933904 2.4565297 4.665674
        sd(sin(2 * pi * Time)) 0.1164408 0.8521297 6.236003
        . . .