Very brief re-post, from March 18, 2006 - now with a little more added commentary:
I remember from an old review that John Palmer did a study on the diurnal pattern of copulation in humans some years ago. You can see the abstract here.
Now, Roberto Reffinetti repeated the study and published it in the online open-access Journal of Circadian Rhythms here:
The two studies agree: The peak copulatory activity in people living in a modern society is around midnight (or, really, around bedtime) with a smaller secondary peak in the morning around wake-time. This makes sense, as natural (pre-Edison) pattern of human sleep is bi-modal: two bouts of sleep. One bout starts at dusk. The second bout ends at dawn. And there is not much to do for a couple of hours of wakefulness in the middle of the night. You can stand sentry. You can think deep philosophical thoughts. Or, if you are there with your partner...well, you know what to do.
Dig through the papers yourself for additional data on workday-weekend differences and the temporal patterns of the female orgasm.