[PLASKETT] THE SPECTROSCOPIC BINARY ,? ORIOXIS 221 



It was not difficult to trace periodic clianges in the velocities thus 

 determined and comparatively early in the present season the period was 

 found to be very nearly 21.90 days. The Potsdam observations, how- 

 ever, did not group themselves satisfactorily with this period and owing 

 to their probably inferior accuracy were not considered. 'The Yerkes 

 observations showed a fairly satisfactory agreement although there 

 were some discrepant single plates due possibly to accidental errors of 

 setting on the rather broad lines of the spectrum or to another cause to 

 be referred to later on. The Lick observations, extending over seven 

 years, followed the velocity curve determined as closely as could be ex- 

 pected although, as there are only five plates, this agreement may be 

 accidental. It was found, however, that a period of 21.87 instead of 

 21.90 days was required to bring the Lick observations forward to ours. 



Although some discrepancies are to be expected on account of the 

 small range in velocity and the relatively high errors of measurement, 

 still it was felt that all the irregularities noted could not be explained 

 on the above grounds. Consequently although sufficient evidence had 

 been secured of the binary character of /? Orionis and sufficient data 

 to obtain the elements of the orbit by the end of January, it was deemed 

 desirable to contimie the observations in the hope of finding a clue to 

 some of the anomalies. The later observations revealed some peculiar 

 and interesting features in the star's motion which served, if not to ex- 

 plain the cause of the irregularities, at least to indicate a reason for their 

 existence. 



The phase of minimum velocity due January 30-31 followed predic- 

 tion, but the succeeding maximum, due Febniary 6-7, although present, 

 was of much lower amplitude than those previously obtained. The curve 

 already drawn showed a range of velocity between about +17 and + 29 

 km. The maximum of February 6 reached only about 23 km. and the 

 succeeding minima and maxima until the end of the observations were 

 as follows:— +19.5, +23.0; +18.5, +24.5; +19.0, +24.5. All of these 

 values as well as the previous ones depend upon several plates and there 

 is no doubt in my mind that they indicate, if not a change in the ampli- 

 tude of the velocity curve, certainly some progressive shift in the position 

 of the absorption maximum of the lines measured due to some physical 

 cause in the star's atmosphere. If it is a change in the amplitude of the 

 motion it may be due to the presence of a third body and will probably 

 be periodic. If an epoch of low amplitude occurred in 1901-1902 this, 

 together with the fact of their only making one plate per night and the 

 consequently higher accidental errors, would form a sufficient explana- 

 tion why Frost and Adams with the high accuracy of their work were 



