358 TBANSACTIONS OF SECTION A. 



By spectroscopic analysis the stars may be classified according to their 

 physical characteristics, and their velocities in the line of sight may be deter- 

 mined. For the purposes of classification, objective priems have been generally 

 employed, and the spectra of many stars obtained on one photographic plate. 

 The measurement of velocities in the line of sight is a more delicate operation, 

 and although initiated in the 'sixties, it was not till the application of photo- 

 graphy thirty years later that results of value were obtained. This class of 

 observation requires a large telescope and a spectroscope very carefully 

 designed and constructed. 



This very brief summary of the different kinds of observations made in the 

 study of the stars may remind you to what extent progress has been dependent 

 on the development of astronomical instruments. The desire to examine fainter 

 objects, and still more the necessity of increasing the accuracy of observations, 

 has brought about a continuous improvement in the range and accuracy of 

 astronomical instruments. Methods which had been perfected for observations 

 of a few stars have been extended so that they can be applied to a large 

 number. For these reasons the progress of Sidereal Astronomy may seem to 

 have gone on slowly for a time. The more rapid progress of recent years 

 arises from the accumulation of data, for which we are indebted to generations 

 of astronomers, and from the gradual increase in power and perfection of our 

 instruments. 



The first insight into the stars as a whole naturally came from the survey 

 of their numbers and distribution ; and Herschel, who constructed the first 

 great telescopes, explored the heavens with untiring skill and energy, and 

 speculated boldly on his observations, is justly regarded as the founder of 

 Sidereal Astronomy. In his great paper, ' On the Construction of the Heavens,' 

 Herschel gives the rules by which he was guided, which I should like to quote, 

 as they may well serve as a motto to all who are engaged in the observational 

 sciences : 



' But first let me mention that if we would hope to make any progress in an 

 investigation of this delicate nature we ought to avoid two opposite extremes 

 of which I can hardly say which is the most dangerous. If we indulge a 

 fanciful imagination and build worlds of our own, we must not wonder at 

 our going wide from the path of truth and nature; but these will vanish like 

 the Cartesian vortices, that soon gave way when better theories were offered. 

 On the other hand, if we add observation to observation, without attempting 

 to draw not only certain conclusions but also conjectural views from them, 

 we offend against the very end for which only observations ought to be made. 

 I will endeavour to keep a proper medium ; but if I should deviate from that 

 I could wish not to fall into the latter error.' In this spirit he discussed the 

 ' star gauges ' or counts of stars visible with his great reflector in different 

 parts of the sky, and concluded from them that the stars form a cluster which 

 stretches to an unknown but finite distance, considerably greater in the plane of 

 the Milky Way than in the perpendicular direction. He gave this distance as 

 497 times that of Sirius. He did not hesitate to advance the theory that some of 

 the nebulpe were similar clusters of stars, of which that in Andromeda, judging 

 from its size, was the nearest. Herschel had no means of telling the scale of 

 the sidereal system, though he probably supposed the parallax of Sirius to be 

 of the order of l". 



Though some of the assumptions made by Herschel are open to criticism, 

 the result at which he arrived is correct in its general outline. I shall attempt 

 to give a brief account of some of the principal methods used to obtain more 

 definite knowledge of the extent and constitution of this ' island universe.' 

 The stars of which most is known are, in general, those nearest to us. If the 

 distance of a star has been measured, its co-ordinates, velocity perpendicular 

 to the line of sight and luminosity are easily found. In the case of a double 

 star whose orbit is known the mass may also be determined. But only a very 

 small proportion of the stars are sufficiently near for the distance to be determin- 

 able with any accuracy. Taking the distance corresponding to a parallax of 

 1" or the parsec as unit — i.e., 200,000 times the distance of the Earth from the 

 Sun — fairly accurate determinations can be made up to a distance of 25 parsecs, 

 but only rough ones for greater distances. 



