Datasets     HST start      mean leg temperatures   Humidity
                               start   middle   end       %
 ( inc990321      11:50        -0.8     0.3     1.0      90 )
   inc990403      14:18         2.7     2.5     3.0      70
   inc990404      12:50         2.7     3.6     4.8      50
   inc990409      14:20         2.3     2.6     3.0      50
   inc990410      14:25         1.0     2.1     2.9      70
   inc990411      10:10        -1.6    -0.6     0.3      90

The difference between the model derived from the latest data (990411) and the previous model (of 990321) is shown below :

The agreement is good, and the rms scatters in the (F1,F2,F3) differences are fairly small, at (0.40", 0.29", 0.33"). Intermediate differences are of a similar nature.
The symmetry remains good (see the plot ). The new track model was installed at 15:00 HST on 19990416.

Strain Gauge Data
The new 'inclinometry' acquisition software now includes strain gauge data as described earlier . The azimuths at which these data (and the inclinometry data) are acquired were adjusted on the basis of the earlier work, so as to better catch the peak strains. Two load measurements (G1,G2) are made of the inner and outer parts of the track. G1 peaks at azimuths 104.54 (wheel 4), 187.74(3), 284.54(2) and 367.74(1), while the G2 peaks occur 0.2 degrees later. The G2 results also have a sign difference of no import. The results below are baseline corrected, and are given in raw units :

         wheel        4           3           2          1
                   -------     -------     -------    -------
    dataset        G1   G2     G1   G2     G1   G2    G1   G2   
                   -------     -------     -------    -------
   inc990403      114  -17    299  -48    215  -45   132  -10
   inc990404      104  -16    297  -48    204  -45   129  -11
   inc990309      105   -4    249  -29    167  -37    94   -5
   inc990311       66   -6    253  -34    187  -39   103   -4

(The 990410 data were taken with the old software - i.e. inclinometry only).
If no additional calibration is necessary, these data indicate considerable variation, with a general decline in apparent load from a (G1) total of 760 units on 990403 to 609 on 990411 : a decrease of 20%.