Added on Nov 3 2008.
Note of March 27, 2011: the content of this page has been temporarily obscured (white text on white background), because it needs revision. If you want however to read it, please "select all" and you will see the text in negative.
More notes on my 1999 tests.
On Aug 11 1999 I was taking the known stationary pendulum recordings (12h centered on the eclipse).
On Aug 12 morning, before accessing the stairway (unaccessed since the evening of Aug 10) to dismantle the setup, I had a look at the monitor (which was on the upper floor along with the VCR) and noticed that the pendulum was visibly out of place. Although I thought of a problem in the structure, I decided to take an unplanned recording, for future review.
Then I dismantled the pendulum.
After that, and also having noticed the strange behavior in the eclipse day, I decided to make other tests to get control data and to check whether there were problems in the structure, or thermal issues.
Particularily, I took recordings at the new moon of September 1999.
I observed a phenomenon that could be "figuratively" described as a push to the pendulum from the sun/moon system, in inverse proportion to the sun-moon angular distance.
Taking as a reference the vertical as shown by the pendulum on September 8 at 05:00 UT (06:00 local solar time, a bit after sunrise), it happened that:
Sep 8, 05:00 UT, sun-moon angle = 21 degrees.....reference vertical as shown by the pendulum
Sep 9, 05:00 UT, sun-moon angle = 8 degrees......the pendulum moved toward about NW
Sep 10, 05:00 UT, sun-moon angle = -5 degrees......the pendulum moved toward about NW even more.
Note that the absolute angle was respectively 21, 8 and 5 degrees, that is decreasing, while the tilt was increasing.
Hence I reinterpreted the Aug 12 strange observation. It happened that:
Aug 11, eclipse.......................................................reference vertical as shown by the pendulum
Aug 12, 10:48 UT, sun-moon angle = -13 degrees......the pendulum moved toward about SE
The absolute angle was respectively 0 (eclipse) and 13 degrees, that is increasing, and, according to what observed one month later, the vertical should have moved toward about SE as compared to the vertical of the day of the eclipse, and this is what happened.
The magnitude was in the order of 10e-5 rad, too much for tidal effects or crust tides.
An unlikely gravitational pull on the building would have shown deviations opposite to the observed ones.
Thermal effects would be easily ruled out because the weather conditions were quite stable in each series of tests, and the observations were made at the same hours in September 8, 9 and 10, and at the same hours in August 11 and 12.
It seemed to me that this phenomenon is observable only for relatively small sun-moon angles, say below 20 degrees.
In the illustrative and widely exaggerated sketch the black pendulum is the reference vertical as in Sep 8 and Aug 11, the red pendulum is the vertical in Sep 9, the blue pendulum is the vertical in Sep 10, and the green pendulum is the vertical in Aug 12. Note: this sketch wan't mean that the reference vertical was the same in August and September, it is only illustrative.
I observed this phenomenon several times with my second pendulum too (the one pictured on the home page).