[molpro-user] Imaginary Vibration Number problem

[MHT]

Hi, 
I want to optg a TS and do path.

After the TS optg was finished, I did a "FREQUENCIES CALCULATION" and got 
the following result:
 Imaginary Vibration  Wavenumber
        Nr             [1/cm]
        1              433.47
        2                9.94
        3                4.65
        4                0.27
                                                                                                      
   Low Vibration      Wavenumber
        Nr             [1/cm]
        1                0.13
        2                0.55
        3                9.27
                                                                                                      
     Vibration        Wavenumber
        Nr             [1/cm]
        1              337.56
        2              524.03
        3              790.54
        4              978.88
        5             1038.20
        6             1152.69
        7             1255.09
        8             1369.59
        9             1508.53
       10             1530.86
       11             3077.24
       12             3110.94
       13             3161.49
       14             3341.03

There are four Imaginary Vibration. In classroom, I was told that the 
small imaginary vibrations usually came/originate from algebraic method of 
the calculation-package, and usually could be negligent. 
However, when I go on further path calculation, molpro told me that " 
Combined Powell-Murtagh-Sargent Update of Hessian
                                                                                                      
 Quadratic Steepest Descent -  Reaction Path Following using updated 
Hessian
                                                                                                      
 Hessian eigenvalues:    -0.035860  0.007064  0.048450  0.111188  0.135029  
0.218905  0.274357  0.320069  0.332180  0.343727
                          0.351616  0.363499  0.387132  0.394537  0.618873
                                                                                                      
 IDIR=-1 requested but starting point is not a stationary point. Gradient 
norm=  0.78D-04  Step length=  0.30D-02
 ? Error
 ? Not a stationary point
 ? The problem occurs in QsdPath2
 CURRENT STACK:      MAIN     "


Does this mean that molpro cannot accept/tolerate the three small 
Imaginary Vibration?
By the way, where these three imaginary vibration came from and how to 
avoid it?
Thanks!

yours sincerely
MHT