WITH A FOURTEEN FEET REFLECTOR. " 183 



field; and at each new bearing of the telescope, preparatory to a fresh run, the 

 circle of position was altered accordingly. But in this way many measures 

 were necessarily rejected. The farther end of the telescope was steadied by 

 contact with a double opening window-shutter, which gradually yielded to the 

 azimuthal motion, yet with sufficient resistance to maintain the stars in the 

 field at perfect rest. 



In declination, on account of the defects before mentioned, and others of 

 minor importance, more than one measure could seldom be obtained during a 

 single run across the field, and that not cotemporaneously with the transits. 

 In addition to this, the frequent necessity of entire disadjustment of the micro- 

 meter, in order to read off the scale within the field, caused a great waste of 

 time; so that the measurement of the stars in the nebulae, though of less im- 

 portance than the observations with the reflector, occupied very much more 

 time and labour. And it will not be surprising, after this very incomplete 

 statement of the difficulties which it was necessary to encounter, and of the 

 numerous sources of error thence arising, if the following measures should, in 

 a few cases, offer somewhat greater discordances than those which occur in 

 observations with a steady instrument, moving in the parallel of the star, and 

 employed in measuring objects that can bear direct vision and full illumi- 

 nation. 



23. Reduction of the Measures. The value of the screw, since the field would 

 scarce include the sun's diameter, was determined entirely by passages of 

 known stars over a given interval, a number of which were observed nearly 

 every evening on which any measures were taken. It was subject to a very 

 slight fluctuation, depending on changes of temperature and minute differences 

 of focal adjustment; and hence it was checked by a comparison of the measures 

 of the largest differences of declination on different evenings. One revolution 

 of the screw at its mean state was equal to 17".640. 



The zero error was always measured by daylight, and was applied as a cor- 

 rection to the intervals of passage on all stars which were taken to determine 

 the value of the scale, and to all measures in declination. During the fall 

 months it increased gradually and uniformly from -f- 1".57 to + 3 ".83. 



There is evidently no allowance necessary for the thickness of the wires, on 

 account of their partial illumination, in any of the different uses to which the 

 micrometer was applied in the present instance. 



