date error occurs during April 2004

[Stupendous Man <richmond@stupendous.cis.rit.edu>]

  Following Gary Billings' idea of using asteroid positions
to check the date in Mark IV data, and Richard Miles' suggestion
of 739 Mandeville during April, 2004, I find the following:

        real date and time                  Mark IV date and time

    Apr  2 UT (JD 2453096.86)       Mark IV position matches JPL ephemeris

    Apr 14 UT (JD 2453109.82)       Mark IV position from "JD 2453108.82"
                                            matches JPL ephemeris
  
    Apr 15 UT (JD 2453110.83)       Mark IV position from "JD 2453109.83"
                                            matches JPL ephemeris
    

    Apr 26 UT (JD 2453120.79)       Mark IV position matches JPL ephemeris



  In all these detections, the Mark IV database showed detections
at the asteroid's position only on the given date and time; in other
words, I wasn't confusing ordinary stars for the asteroid.

  So, I conclude that the Mark IV database has an "off-by-one-day"
error for _part_ of the month of April, 2004.  This suggests to me
a problem with the clocks on the computers, rather than with the software
which calculates JD. 

  I need to write a script to go through _all_ the Mark IV data
and search for asteroids in order to see if this is the only
date-and-time problem in the database.  Michael Sallman's
database interface will help greatly.

  More later.

                                             Michael Richmond