OF ARTS AND SCIENCES. 385 



TABLE YII. — LiGHT-CuuvE of DM. 8r25. 



From the last three gi'onps we may infer that the light of this star, 

 like that of the others of the same class, is constant except during the 

 few hours immediately preceding or following the minimum. 



The same precautions were taken as with fi Persei in drawing the 

 lisht-curve that it should be free from sudden changes in curvature. 

 From the small residuals in the last column we may therefore infer 

 that the accidental errors are very small. About an hour before the 

 minimum the light ceases to vary, and remains nearly constant for an 

 hour and a half, when it begins to rapidly increase. The exact time 

 of these changes may be found more precisely by subdividing the 

 groups whose means are — 89 and -[-60. Making the period of the 

 groups ten minutes instead of thirty, we replace the first group by 

 three containing 4, 5, and 5 sets, having mean times 99, 90, and 80, 

 and magnitudes 1.54, 1.30, and 1.21. The other group similarly sub- 

 divided gives for the mean times 52, 60, and 69, the magnitudes, 

 1.25, 1.26, and 1.44. From these we might infer a somewhat longer 

 period of uniform light than would be indicated by the curve already 

 drawn. The number of observations is, however, too small to deter- 

 mine this point with certainty. 



The correction to the ephemeris of the minima is best found by 

 Table VIII., which gives the time at which the light is equal while 

 VOL. XVI. (n. s. VIII.) 25 



