VAJ; TABLE STAKS. 



2.— VARIABLE STAKS OF LONG PEKIOD. 



By PROFESSOR E. C. PICKERING, Astronomical Observatory, 

 Harvard University, U.S.A. 



The astronomers of Australia have one great advantage over 

 those in Europe and the United States. F^w observations^ have been 

 made in a large part of the sky within their reach, the region south 

 of declination — 30 deg. This is particularly the case with one of the 

 most interesting and fruitful fields of work in astronomy, the study 

 of the variable stars. The method described below has been used very 

 extensively for the northern stars at the observatory of Harvard 

 College. It has the great advantage that no instrument is required 

 but a telescope, and little skill beyond that needed to identify with 

 certainty a faint star. Without the latter ability, much useful work 

 on the stars cannot be done. For the northern stars, the atlas of the 

 Bonn Uurchmusterung is almost indispensable. A similar atlas for 

 the southern stars, including those of the ninth magnitude, and 

 bi'ighter, is still one of the greatest needs of astronomy. 



The variable stars of long period are objects of special interest. 

 At maximum, they are many times as bright as at minimum. Their 

 changes are irregular, or at least unknown. It is, therefore, impor- 

 tant that they should be observed at frequent intervals to enable 

 astronomers, in the future, to determine the laws regulating their 

 changes. The method adopted at the Harvard Observatory for 

 observing these stars is as follows : — A list was first prepared of all 

 the variables brighter than the ninth mag-nitude at maximum, and 

 undergoing changes of more than three magnitudes. A sequence of 

 comparison stars was next selected, each being about a third of a 

 magnitude fainter than the next, the brightest being brighter than 

 the variable at maximum, and the faintest being fainter than the 

 variable at minimum. The magnitude of these stars is then deter- 

 mined photometrically, with great care. A photograph of the region, 

 taken with the 24-in. Bruce telescope, and having an exposure of an 

 hour, is then enlarged three times on an 8 by 10 plate. Prints from 

 these enlargements gives maps of regions a degree square, on a scale 

 of 20 seconds to a millimetre. The magnitudes of the stars are 

 entered on these charts, omitting the decimal point, so that it shall 

 not be mistaken for the image of a star. Thus, magnitude 11'2 is 

 written 112. To observe the variable, it is only necessary to identify 

 the region seen in the telescope with that of the chart. Thus, if the 

 variable is fainter than a star marked 95, and brighter by an equal 

 amount than a star marked 99, its magnitude will be 9 '7. If more 

 nearly equal to the first star, it will be 9'6, &c. Copies of several of 

 these charts are enclosed, and may be kept by any astronomer here 

 present who is ready to imdertake the observations. The results may 

 be plotted directly, using days for abscissas and magnitudes for 

 ordinates. The observer should not know in advance the probable 

 magnitude of the star, as his observations will have very little value 

 unless they are independent. About 300 variables north of declina- 

 tion 30 deg. south are now being followed in this way at Harvard. 



