October 20, 1911] 



SCIENCE 



525 



redness, the average density for types G and K 

 being about that of the sun. 



8. It may be added that all these facts (except 

 the existence of one very faint star of type A) 

 are in harmony with the scheme of stellar evolu- 

 tion sketched by the writer last year. 

 The 6-inch Transit Circle of the U. S. Naval Ob- 

 servatory: F. B. LiTTELL. 



Work has been begun with this instrument on a 

 list containing the old or historical fundamental 

 stars, and the new fundamental stars as proposed 

 by the International Committee of the Photo- 

 graphic Chart of the Sky. The instrument is 

 fairly satisfactory as to its errors, which are either 

 quite constant, or are readily determined for any 

 instant by interpolation between values observed 

 at suitable intervals. The pivots have been meas- 

 ured with great accuracy by the axial microscope 

 method. Their irregularities are very small. The 

 division errors of the degree lines of circle A 

 have been determined, also the periodic errors of 

 the two-minute lines within the ten-minute spaces. 

 The periodic and progressive errors of the tele- 

 scope micrometers and of the microscope microm- 

 eters have been measured. 



The zenith distance micrometer is used for 

 declination work. Readings are made on two circle 

 divisions under each of the four microscopes, using 



and reflected (D.E.), clamp east and west (E.W.), 

 and with objective and eye-end interchanged 

 (I., II.). Circle A was used in fixed position, and 

 circle B was shifted each night. The level errors 

 were determined from nadir observations. 



Eight Ascension. — By a discussion of the dif- 

 ferences B.-W. and R.-D. the following flexures 

 were wound: 



Lateral flexure varying as cos 2z 



+ 0».005 ± 0».0Ol (DEI) 

 Axis flexure varying as cos 2z 



— 0».025 ±0^002 (DEI) 



The following tables show the quantities indi- 

 cated before and after the application of the cor- 

 rections due to the above flexures. 



In the next table each quantity is the mean for 

 3 stars and each has been multiplied by cos 5. 



two pairs of threads in each, distant 2.5 revolu- 

 tions of the micrometer screw from each other. 

 A hand-driven self -registering right ascension 

 micrometer is used. A reversing prism is used, 

 by means of which each observation is taken half 

 with image direct and half with image reversed. 

 The instrument is reversed from clamp east to 

 west and vice versa at short intervals. 



In order to test the performance of the instru- 

 ment a list of 30 stars was observed 4 times in 

 each of 8 positions of the instrument; i. e., direct 



By the application of the flexure corrections the 

 probable error of a single observation in right 

 ascension is reduced from 0.0266 see, 5 to 0.0204 

 sec. 5. For direct observations it is reduced to 

 0.0190 sec. 5. 



The above error is entirely eliminated from the 

 mean of an equal number of clamp east and clamp 

 west observations. 



The following table gives the comparison with 

 the right ascensions of Newcomb (N), and those 

 of Boss (B), tabulated as above. 



