PERIODICAL STARS. 223 



and therefore unmeasured and probably undeterminable periods, 

 or which, in a few, occur without being periodical, as it were, 

 by a sudden revolution, either for a shorter or for a longer 

 time. The latter class of phenomena (of which a remarkable 

 instance is furnished in our own days by a large star in Argo) 

 will not be here discussed, as our proper subject is those 

 fixed stars whose periods have already been investigated and 

 ascertained. It is of importance here to make a distinc- 

 tion between three great sidereal phenomena, whose con- 

 nexion has not as yet been demonstrated; namely, variable 

 stars of known periodicity; the instantaneous lighting up 

 in the heavens of so-called new stars; and sudden changes 

 in the luminosity of long-known fixed stars, which pre- 

 viously shone with uniform intensity. We shall first of all 

 dwell exclusively on the first kind of variability ; of this the 

 earliest instance accurately observed is furnished (1638) by 

 Mira, a star in the neck of Cetus. The East-Friesland 

 pastor, David Fabricius (the father of the discoverer of the 

 spots on the sun), had certainly already observed this star 

 on the 13th of August, 1596, as of the 3rd magnitude, and 

 in October of the same year he saw it disappear. But it 

 was not until forty-two years afterwards that the alternating, 

 recurring variability of its light, and its periodic changes, 

 were discovered by the Professor Johann Phocylides Holwarda, 

 Professor of Francker. This discovery was further followed 

 in the same century by that of two other variable stars /3 Persei 

 (1669), described by Montanari, and x Cygni (1687) by Kirch. 

 The irregularities which have been noticed in the periods, 

 together with the additional number of stars of this class which 

 have been discovered have, since the beginning of the nine- 

 teenth century, awakened the most lively interest in this 

 complicated group of phenomena. From the difficulty of 

 the subject, and from my own wish to be able to set down 

 in the present work the numerical elements of this variability 



