4 THE CORRECTION OF SEXTANTS FOR ERRORS 



plane of the arc can then be brought parallel to that of the circle by 

 means of the table-screws. The arc now rotates in its own plane when 

 the table revolves, and the extent of any such motion can be accurately 

 measured by the verniers of the circle. The index-bar being in any })o- 

 sition desired, with the direct view through the horizon-glass cut off by 

 interposing the colored screens, let the table be turned until the i"efiected 

 image of the colliraator-mark is bisected by one of the telescope wires, 

 and let both sextant and circle be read. Now move the iudex-bar into 

 any other position, bring the mark back upon the same wire by turning 

 the table in the opposite direction, and again read the sextant and circle. 

 In this process the angle of rotation has obviously been measured by 

 both instrument-', and the difference betweeu the two results is, therefore, 

 the correction of the sextant for that angle, if the readings of the circle 

 have been duly corrected for their own errors. The circle referred to has 

 never been thoroughly investigated, but a preliminary examination showed 

 that its errors are probably small ; a double axis permits any part of the 

 graduation to be used, and the effect of eccentricity is, of course, always 

 eliminated. Except the circle and its immediate appendages, Avhich were 

 taken from a theodolite of exquisite workmanship by Gambey, of Paris, 

 this apparatus, intended rather to illustrate a j)roposed system of exam- 

 ination than for service, was roughly made, but it proved to be efficient, 

 and was sent to the Naval Observatory some years afterward, where it 

 is, I believe, still occasionally used. 



As compared with Mr. Cooke's system of collimators the apparatus 

 just described presents some important advantages. Direct reference to 

 the circle which serves as a standard is preferable to a comparison 

 depending upon the permanence of intermediate arrangements, even such 

 as are believed to be of a stable character. Instead of being restricted 

 to a few angles, which it must be inconvenient, at least, to change, any 

 part of the graduation may be tested at pleasure ; the position of every 

 line on the arc can be verified if that is desired ; and the portability of 

 an apparatus which can be moved from one room to another by a couple 

 of persons is a minor point in its favor. The collimators, on the other 

 hand, conveniently dispense with all readings except those of the sextant 

 during the examination. 



Before commencing the examination of a sextant, the principal adjust- 

 ments must be tested, and rectified if necessary ; both mirrors must be truly 

 parallel to the axis, and the line of coUimation parallel to the plane of the 

 limb ; the state of the index correction is immaterial. The telescoj^e of 

 the sextant may be used when of sufficient power, but can advantageously 

 be replaced by a special one, magnifying at least ten diameters, with a fine 

 wire in the middle of the field crossed exactly at right angles by two other 

 wires which inclose a central space of convenient width. 



J 



