The following changes are required to the first printing of the
Amber 4.0 manual. (Later printings are numbered on the second page.)

************************Minmd section, page 81, item 11.4

************Original

11.4   DIELC      Dielectric multiplicative constant for the electrostatic
         interactions.  If DIELC .le. 0.0 then DIELC = 1.0.  DIELC
         and IDIEL are coupled.  For example to obtain a dielectric
         constant of 4rij set DIELC=4 and IDIEL=1.

************Should read

11.4   DIELC      Dielectric multiplicative constant for the electrostatic
         interactions.  If DIELC .le. 0.0 then DIELC = 1.0.  DIELC
         and IDIEL are coupled.  For example to obtain a dielectric
         constant of 4rij set DIELC=4 and IDIEL=0.
						^
						|

************************Gibbs section, page 105, paragraph 1

Original

    GIBBS 4.0 is built upon the  original  (3.0)  version  of  this
    code,  which was written by U.C. Singh and P.A. Kollman (UCSF),
    using numerous MD routines adapted from GROMOS83  by  W.F.  van
    Gunsteren.  

Should read (additional credit)

    GIBBS 4.0 is built upon the  original  (3.0)  version  of  this
    code,  which was written by U.C. Singh and P.A. Kollman (UCSF),
    using numerous MD routines adapted from GROMOS83  by  W.F.  van
    Gunsteren.  Speedups  to  non-bonded  pairlist  generation  and
    residue-based imaging  adopted  from  code  written  by  George
    Seibel for AMBER revision 3A.

************************Gibbs section, page 120

************Original

10.2   NTID       Read, but not used.

************Should read

10.2   NTID       Flag for solvent pairlist behavior.

        = 0  only the first atom of each solvent molecule is used
             when generating the non-bonded pairlist for a periodic
             system (for water, this is the oxygen). If this atom
             lies within the specified cutoff, the entire solvent
             molecule is included in the non-bonded pairlist.  This
             can result in a substantial speedup in non-bonded
             pairlist generation, and is recommended when using water
             as the solvent.

        =86  all atoms in a solvent molecule are considered when
             generating the non-bonded pairlist for a periodic system.
             If any atom of the solvent molecule lies within the
             specified cutoff, all atoms of the solvent molecule will
             be included in the non-bonded list. This is the behavior
             of versions of AMBER <= 3.0.

         A value of NTID=0 is suggested for calculations using water as
         a solvent. For calculations using larger solvent molecules, one
         should carefully consider whether using only the first atom is
         appropriate.

         NTID will have no affect for non-periodic systems.

************************Gibbs section, page 130

**********Original

14a.3  ISLP       Determines the direction in which the slope is calculated.
        
        = 1  the slope is calculated in the forward (0->1) direction
        =-1  the slope is calculated in the reverse (1->0) direction
        = 2  the slope is calculated using the average of the forward and
             reverse energies.

         For best results, the slope should be calculated in the same
         direction as the simulation, or as the average of the forward and
         reverse energies. The default behavior (ISLP=0) is to calculate the
         slope in the same direction as the simulation.

         With thermodynamic integration, or when double-wide sampling is
         defeated, ISLP has no effect.

**********Should read

14a.3  ISLP       Determines the direction in which the slope is calculated.
        
        = 0  (default) use the appropriate value of ISLP (-1 or 1) to 
             calculate the slope from energies calculated in the same
             direction as the simulation (recommended).
        = 1  the slope is calculated from the forward (0->1) energy
             at each step.
        =-1  the slope is calculated from the reverse (1->0) energy
             at each step.
        = 2  the slope is calculated using the average of the redundant
             free energy values (from double wide sampling) over the 
             interval in the direction opposite to the simulation, i.e.
               G(reverse[curr window] - G(forware[prev window])/2 or
               G(forward[curr window] - G(reverse[prev window])/2 
             for simulations run 0->1 and 1->0, respectively.
             This option can be useful when very few points are used
             to evaluate each slope (e.g. IAVSLP = 2).
        = 3  the slope is calculated using the average of the forward and
             reverse energies at each lambda.

         For best results in most cases, the slope should be calculated
         in the same direction as the simulation. This is the default
         behavior (ISLP=0).

         With thermodynamic integration, or when double-wide sampling is
         defeated, ISLP has no effect.

         Only options ISLP=0 or ISLP=3 should typically be used when 
         AMXRST > 0.