NavList:

From: Robert VanderPol II <<Have you looked into Chip Scale Atomic Clocks?  Somebody mentioned $1500 but I don't know for sure.  Don't know what the power requirements would be either.>>

You can buy the devices themselves.  I was unable to find products intended for timekeeping.

From: Peter Monta <<excellent lengthy post>>

Thank you for the background, which I found insightful.  I'll look around for surplus products.  A problem I have encountered is that OXCOs etc are more readily available as frequency references than as timekeeping products.  Building a timekeeping device around a frequency reference is a science project that I am not inclined to pursue.

From: John D. Howard <<

I do not understand the need for super acurate watches.  Any quartz watch will keep time for a month or so with reasonable acuracy.  Unless you plan to be at sea many months without making landfall ( two women and two dogs ? ) than you can check your watch, to the second, using your sextant. You will have paper charts that have lat and long for most places.  With that and your sextant ( and almanac ) do a time sight.  Set your watch or note how far off it is, down to the second.  No radio, internet, or GPS needed. Like Frank said - invert it.  Use cel nav to get the time!>>

You are, of course, correct that the necessary accuracy is only that needed to find a reasonably accurate latitude when making landfall at the end of the longest passage in the voyage, presuming that reasonably accurate charts are available for the area from which the passage begins.  Recreational vessels rarely make passages lasting longer than 30 days, so the question becomes whether the error in a cheap quartz watch is acceptable after 30 days.  With a good quality mechanical watch, you could be off by 150 seconds, with a cheap quartz watch, you could be off by 30 seconds.  Maybe that's close enough.  I am a novice, but it seems like 30 seconds of avoidable uncertainty is worth thinking about, at least a litte.