178 CARNEGIE INSTITUTION OF WASHINGTON. 



this mean is 6 kilometers. The deviation of the individual values of 

 space velocity from the mean is no larger than in the case of the mem- 

 bers of the Taurus group. The seven members of the group are widely 

 scattered over the sky. 



Another significant feature of the investigation results from a dis- 

 cussion of the space velocity arranged according to type. Starting 

 with a mean velocity of about 40 kilometers for M and K type "giants," 

 the velocity by type decreases to about 15 kilometers for stars of type 

 B and then rapidly increases for "dwarf" stars, apparently reaching 

 a maximum about the G-type. The phenomenon is easily explainable 

 on the assumption that only the stars of greatest mass reach the 

 B-type development. 



While the investigation is in shape for printing, the publication will 

 be delayed awaiting the revision of the system of parallaxes through 

 the acquisition of additional data which has been generously offered 

 by Professors Schlesinger, Mitchell, Miller, Campbell, and Frost. 

 Professor Pickering has made it possible to revise the study of the 

 absolute magnitudes by supplying the spectral types as determined 

 for the new Draper Catalogue. 



PERSONALITY IN THE ESTIMATION OF TENTHS. 



In estimating the position of an index between two scale-rulings to 

 tenths of the scale interval personality plays an important part. 

 The list of percentages giving the frequency with which each tenth 

 is represented in the estimations of an observer is known as the 

 observer's personal scale. In connection with the design of measuring 

 scales for a special measuring-instrument it seemed desirable to know 

 the magnitude and nature of errors to be anticipated due to the per- 

 sonal scales of the measures. Accordingly, Dr. Albrecht has made a 

 detailed investigation of the personal scales for sixteen present and 

 former observers of the Dudley Observatory, extending over periods 

 of more than twenty years for some of the observers. The principal 

 results may be briefly summarized as follows: 



(a) The personal scales for different observers may differ widely 

 from each other. 



(b) For any particular observer the personal scale is remarkably 

 constant over short and moderately long periods of time, even up to 

 one year. 



(c) Fatigue appears not to change the personal scale appreciably. 

 For example, it remains essentially constant during 5 hours of con- 

 tinuous observing at great speed. 



(d) For some observers the personal scale remained practically un- 

 changed for 8 or 10 years, the entire period covered by the data, while 

 for other observers it was found to undergo a gradual and more or 

 less progressive change from year to year. 



Tests were also made on specially ruled scales in order to determine 

 the extent to which different widths of the scale-rulings and of the 



