I arrived at USNO on April 1, 1991. The first week was used to set up 
living in Washington. I than began introducing myself to the field of 
optical interferometry and the software run by the MK3 group to reduce 
data from the optical interferomter on Mt. Wilson. I installed a new 
Fortran compiler and adapted our standard graphic software to it. 

After about one month, when I was supplied with a PC, I became involved 
in the reduction of binary star data. As a first exercise, I reduced 
data for Alpha Andromedae, which has been analyzed by X.-P. Pan et al. 
(in prep.), and, in the course of this, I implemented an algorithm to 
fit an ellipse to the apparent orbit of a double star. Thus (by applying 
the Thiele-Innes method), good starting estimates of the orbital 
elements can be derived for subsequent non-linear least squares fitting 
to the data. 

After having gained some experience, I started working on Theta 2 Tauri. 
This is a binary system with a highly elliptical orbit, which takes the 
two components as close as about 4 milli seconds of arc every 141 days. 
At these small separations the relative position angle of the two 
components changes significantly (about 8 degrees) during one night, 
thus invalidating our assumption of a constant angle, whose value (and 
the value of the radius) is derived from the whole night's data. A 
solution to this problem would be a global least squares fit of the 
orbital elements directly to the visibilities. However, I worked out a 
modification to the existing software to cope with a known linear rate 
of change of the position angle, which can be derived from a preliminary 
orbit (which can be determined from nights when the system is at a 
larger separation). 

I spent the last two weeks in June on Mount Wilson (California) for 
observations with the MK3 optical interferometer (jointly operated by 
the NRL and USNO). The seeing conditions were mediocre, although we had 
many clear nights. More than half of the time was lost due to high 
humidity (which jeopardizes our silver mirrors when condensation on 
their surfaces takes place) and technical problems. An earthquake took 
place on early morning of Friday 28. Epicenter was near our location in 
the San Gabriel mountains, about 7 miles below the surface, and 6.0 
on the Richter scale. It lasted about 18 seconds. Our buildings, 
inventory, and the interferometer experienced no damage.