APPENDIX D LXXXV 



trolling circuit, but if it be dosed again without due precautions, if the 

 phase-difference between the synchronizing and synchronized clocks be 

 excessive, there is a risk of stoppage of the latter. This danger has 

 been eliminated by the installation of a safety cut-out in each master- 

 clock, by which, if any accident happens to the line, the clock automati- 

 cally cuts itself out of synchronization, and runs as an independent 

 clock until normal conditions are restored. 



An arrangement has also been introduced by which it is possible 

 to transmit the beats of the signal clock over the telephone line at any 

 time. This is managed by a special transmitter, in front of which is 

 mounted a telegraph sounder which can be operated by the clock. The 

 telephone signals are in frequent requisition by surveyors, jewellers, and 

 others who require exact time. 



Astrophysics. 



The principal work in Astrophysics has been along the line of stel- 

 lar spectroscopy as applied to the determination of the radial velocities 

 of some selected stars. This work has been actively prosecuted during 

 the past year, about 800 negatives of stellar spectra having been made 

 and considerably over half of these measured and reduced. The majority 

 of these spectra have been of stars known to be spectroscopic binaries, 

 and the observations have been applied to the determination of the ele- 

 ments of the orbits. Satisfactory elements have been obtained for three 

 stars, a Draconis, i Orionis and tp Orionis, and provisional elements for 

 two more, Aquilae and 7 Virginis. In obtaining the elements of these 

 five binaries 275 spectra have been employed, of which 107 were required 

 for z Orionis. The very diffuse nature of the lines in the spectrum of the 

 latter rendered the measures uncertain. Moreover, the high eccentricity 

 of the orbit with the consequently abruptly changing form of the velocity 

 curve, and the probable presence of a secondary disturbance, further in- 

 creased the difficulty of obtaining satisfactory elements, and necessitated 

 the use of such a large number of plates. 



Of the other spectrograms some have been of early type stars with 

 diffuse lines, of which one has proved to be a spectroscopic binary. A 

 large number of the measures, however, are of binaries of which there 

 are not yet sufficient observations to complete the orbit, although several 

 are well under way. 



About 70 spectra of J3 Orionis have been measured in an attempt to 

 determine the effect that an increase of the width of slit has on the ac- 

 curacy of the determination of velocity. The results of this investiga- 

 tion, so far as early type stars are concerned, show that, with the single 

 prism spectrograph, dispersion 30 tenth-metres per millimetre at Hy, 

 the slit may be made 0.051 mm. wide without much increase of probable 



