[LEAPSECS] drift of TAI

[Magnus Danielson]

Zefram wrote:

> David Malone wrote:

>> The cueues for this frequency shift look just like a 4 hour phase

>> shift every day.

> 

> Except that it's a continuous `shift', rather than an instantaneous

> four-hour jump.

> 

>>                 Of course, there could be a long-term adaption

>> I guess.

> 

> That's what I'm expecting.  With jetlag you're expected to retain the

> same frequency and shift phase, so there's no adaptation to a different

> frequency.

> 

>> Unless you were living divorced from daylight, it's unlikely this

>> is your natural period.

> 

> I had *some* exposure to the outdoors, but this was only a couple of

> percent of the time, when taking buses between home and campus.  Half the

> time this did not involve sunlight.  At both ends of that journey I

> lived a completely indoor existence, without unobstructed sight of an

> external window.  So desynchronisation seems quite feasible.

> 

> Probably relevant: before the free-running period I was accustomed to

> irregular sleep periods that had little synchrony with the planet.

> Even when I had lectures to go to, I kept completely ad hoc hours.

> In those years I could shift phase to an arbitrary extent within

> two days, by simply staying up until the bedtime of the target phase.

> Nowadays I've become conditioned to working (approximately) office hours,

> and I seem to be much more tied to the regular cycle: I have difficulty

> staying up as much as 24 hours.  When I need to, though, I still break

> phase entirely rather than shift gradually.

> 

>> 	http://www.hno.harvard.edu/gazette/1999/07.15/bioclock24.html

> 

> Interesting work.  This still, like the discredited earlier work, shows

> a biological cycle longer than the natural solar day.  Evolutionarily

> you'd expect us to be built for a shorter cycle, since the Earth rotated

> faster in our evolutionary past.  But the day was only a few minutes

> shorter when our ancestors switched from nocturnal to diurnal behaviour,

> so there shouldn't be much in it.

> 

>> a 28 hr day is long enough to prevent people's hormones getting in

>> sync with the day.

> 

> They're drawing a strong distinction between the hormone-controlled

> circadian rhythm and actual sleep/wake rhythm.  I hadn't thought about

> this before.  On those occasions when I've stayed awake for lengthy

> periods of time, occasionally 60 and once over 70 hours, I've certainly

> had periods of lowered body temperature and lower physical activity,

> matching what they describe for the hormonally-sleepy phase.  I recognised

> it as such at the time.  But the regular 40-hour cycle was different.

> As I recall, I just didn't feel sleepy at all until I'd been up more

> than 20 hours.  The cycle of 24 hours awake and 16 asleep felt like a

> normal circadian rhythm, not overriding the natural state at any point.


No, no, no... with the new SI second based on the observation that a 
second is 10639620104 and 1/6 oscillation from the hyperfine transition 
of Cs-133. 100 ks becomes just a day.

Let's keep the prespective here! Also, 1 meter is the precission 
calibrated distance from my fingertip to my noise, assuming of course 
that I have been given the necessary beers before you are allowed to 
measure. Instability of measures can't be guaranteed otherwise.

Cheers,
Magnus