384 PROCEEDINGS OF THE AMERICAN ACADEMY 



later observations were made through clouds which finally stopped 

 the measurements. On February 17 seventy-five sets or nine hun- 

 dred settings were obtained ; the observations extended from seven 

 o'clock in the evening until the variable had regained its full light, at 

 about half-past two on the following morning. During this time no 

 interval of more than five minutes elapsed during which an observer 

 was not comparing the two images. During most of the time the ob- 

 servers took sets alternately, so that there was only an interval of a 

 few seconds between the sets. On February 22 observations began 

 at half past six and continued until ten o'clock, when they were 

 stopped by clouds. Twenty-six sets were obtained in this time. A 

 long period of cloudy weather intervened, and the next measures 

 were made on March 24. Thirty-six sets were taken through clouds, 

 from quarter-past nine to quarter-past twelve. Owing to the small 

 distance between the stars, no perceptible error seems to be introduiced 

 by these clouds, as long as they are not dense enough to render the 

 stars invisible. Some measures were obtained on March 14, but ap- 

 parently the wrong star was observed. They were stopped by the 

 deposition of dew on the object-glass, which may have caused an 

 error, since the two pencils include different portions of the ob- 

 jective. No use has been made of these observations. On April 3 

 another minimum was observed. Fifty-two sets of six hundred and 

 twenty -four settings were made between seven o'clock and midnight, 

 when the star had recovered its full brightness. Forty-four sets of 

 five hundred and twenty-eight settings were also made on other even- 

 ings to determine the undiminished light of the star. Fifteen of these 

 sets hy Mr. Searle give its light as 3.64 magnitudes brighter than DM. 

 81°26. Sixteen sets by Mr. "Wendell gave 3.59, and thirteen sets by 

 myself gave 3.71. As the evidence of systematic difference is not con- 

 clusive, the mean of all, or 3.64, has been adopted. 



The entire number of measures, not including those of March 14, is 

 273 sets or 3276 settings. 



Table VII., like Table I., gives the results of these measures ar- 

 ranged in groups in the order of times from the computed minimum. 

 The columns give the mean of the times, the number of sets of twelve 

 settings, the mean magnitude, the corresponding magnitude derived 

 from a curve drawn nearly through them, and the difference of the 

 last two columns. Each group extends over thirty minutes, except 

 the first, which extends from — 311 to — 258, and the last three, 

 which include all the measures made when the nearest minimum was 

 more than five hours distant. Their limits are -{-852 to -f-OOl, 

 -1-1350 to -f 1495, and -|-1887 to -|-22f)0 minutes. 



