336 ANNUAL REPORT SMITHSONIAN INSTITUTION, 1964 



the unknown star and reference stars in the preliminary testing of the 

 machines. 



Third, there inevitably would be errors in the setting of the machine. 

 Preliminary estimates indicated, for example, that when images of 

 40-micron diameter were measured with the Mann machine, there was, 

 in the settings, a consistent and repeatable eiTor of 1 micron on the 

 average. 



A further error could be introduced by the camera itself, although 

 it seemed unlikely that the geometry of the Baker-Nunn system would 

 cause any very considerable error of this sort. In any case, it had to 

 be determined whether the image of the star on the film would be of 

 such a magnitude as to introduce a significant variation in the 

 measurements. 



Finally, and most importantly, distortion of the fihn as placed in the 

 machine might introduce a substantial error. That distortion would 

 not be the same in every direction, and therefore positions reduced with 

 linear plate constants might not be reliable. However, over small 

 distances of 1 or 2 centimeters of film, it was expected that fluctuations 

 in the plate-scale would be small, not exceedmg 1 or 2 seconds of arc. 

 The staff devised a simple method of evaluating this problem by meas- 

 uring the same grouping of stars on several separate frames and then 

 studying the residuals and positions from frame to frame. 



By mid-1958, the photoreduction section had developed an efficient 

 system of filing the films, had midertaken the searching of films with 

 microscope viewer, and was planning the techniques for measuring the 

 satellite images on the films. Again, this was a two-fold problem, 

 one of developing appropriate methods, and the other, at the same time, 

 of training personnel to use them. 



Precision reduction of the Baker-Nunn films of artificial satellites 

 began in June 1958, and by the end of September the positions of some 

 69 satellite images had been precisely determined. It was initially a 

 vei*y slow procedure ; a trained operator could measure between four 

 and six satellite images per day with either the Van Biesbroeck or the 

 Maim measuring engine. 



The initial phases of the work had been carried out by Pedro 

 Kokaras, under the immediate supervision of Drs. "Wliipple, Ilynek, 

 and Henize. In October, however, Dr. Karoly Lassovszky, a refugee 

 from Hungary, joined the staff as astronomer in charge of photo- 

 reduction. Mr. Kokaras then served as his administrative and techni- 

 cal assistant and supervised the work of the measurers. 



During the last quarter of 1958, some preliminary evaluations of the 

 two measuring engines were possible. In those 3 months, 94 images 

 were measured on a modified Van Biesbroeck machine, with a mean 

 estimated probable error of 7.4 seconds of arc in right ascension, and 

 5.5 seconds of arc in declination. 



