1.  The definitions are:
        (int opd) = (delayed arm length) - (fixed arm length)
        (int opd) = 2[LD + SD] +C             ... for tel B = delayed.

     Thus at LD = SD = 0 (home) we get
        C = x(del,home) - x(fix,home)

where x is the optical length of the respective arm.
Each instument will have its own value of C depending on the lengths of the optical paths leading to its beam-combination point.

2.  For the IR table the C value is approximately tabulated in my 19 Oct 95 memo; at 38 m baseline it is
        C(IR) = -1814.1 cm.

3.  For the Fluor instrument it is a custom to write the C value as the sum of 2 parts, a magic constant plus a Fluor offset (fixed in value),
     giving
       C(F) = MC + FO

     which at 38 m has values
        C(F) = -1704.3 + 39.65 cm
                = -1664.6 cm.

4.  The difference is
        C(F) - C(IR) = 149.5 cm

     which is due to about 128 cm greater length in the Fluor delayed arm length, plus about 21.5 cm (perhaps 20.3) greater optical length
     in the delayed arm glass fiber in the combination box.

5.  In autocollimation the paths are slightly different, and we found for 38 m baseline in Nov 1999 that
        MC(auto) = -1709.5 cm,

     i.e., about 5.2 cm more negative than the value for a star, giving in general
        MC(auto) = MC -5.2 cm.

6.  At the 21 m baseline we expect (see 19 Oct 95 memo)
        C(IR) = 202.7 cm

    which leads to
        MC = C(F) - FO
               = C(IR) + 149.5 -39.65
               = C(IR) + 109.9
               = 312.6 cm.

7.  In Jan 2000 the stub delay dihedral was moved about 1.5 cm toward the corner, so the fixed arm length is 2*1.5 = 3.0 cm shorter, so
     in Jan 2000 we have
        C(2000) - C(1999) = -x(fix,home,2000) + x(fix,home,1999) = 3.0 cm.

8.  From the above we compile a table of expected results.