> Is there a convention for MJD with a leap second? I think not. See http://www.ucolick.org/~sla/leapsecs/timescales.html Also not that POSIX time_t suffers from the same problem. -ch