32 THE CORRECTION OF SEXTANTS FOR ERRORS 



the length of the vernier does not affect this result, since one reading is 

 always as much augmented thereby as the other is decreased ; but if an- 

 other limb correction is applied, it should be the mean of the corrections 

 for the three points of coincidence. By this artifice each observation is 

 referred to three distinct positions on the arc, with a corres2:)onding dimi- 

 nutiort in the effect of purely local errors. 



When the presence of large errors upon any small portion of the limb 

 is suspected, in consequence of a local injury or otherwise, as many of 

 the lines in this tract may be examined as the nature of the case requires. 

 These comparisons are not to be employed in finding the eccentricity, but 

 must be reduced separately to determine the local errors. 



If it should become generally known that sextants purchased in con- 

 siderable numbers were inspected only at certain points of the gradua- 

 tion, unscrupulous makers, especially those using copying engines, might 

 be tempted to bestow greater care upon the critical divisions than upon 

 the rest of the arc. But the points to be examined may be selected at 

 pleasure, if their number and distance from each other remain unchanged, 

 as has been shown. In arranging a system of examination for any given 

 service, the number of comparisons in a series will naturally depend on 

 the special requirements of the case. A greater number affords a better 

 representation of the entire graduation, while, on the other hand, it extends 

 the time during which errors may be introduced by gradual or sudden 

 changes in the apparatus, or in the sextant itself. Perhaps there can 

 seldom be any sufficient reason for spacing the comparisons more closely 

 than at points 5° apart. 



An alternative form of the apparatus has been devised, having two 

 collimators — one fixed, the other carried by an arm attached to the circle 

 and directed toward the sextant, which is supported by a fixed table 

 immediately over the circle. The two collimator-marks — one seen direct, 

 the other by reflection — are brought into coincidence in the field of the 

 sextant telescope, as in the ordinary use of that instrument. No appa- 

 ratus of this form has been constructed, but the details have been worked 

 out far enough to show that no serious practical difficulty is to be appre- 

 hended. The principal feature of improvement is that nothing will 

 depend upon immobility of the sextant, which lies freely upon its table 

 during the examination and may be removed for inspection at any time. 

 It may be, however, that this advantage is rather apparent than real, for 

 in either case the sextant must be handled with the utmost caution, and 

 experience has not shown that there is any difficulty in preventing an 

 appreciable displacement. On the other hand, the proposed form must 

 necessarily be more expensive than the existing one, the illumination is not 

 so easily effected, and probably the observations will not be quite so good 



