1835.] Collimation Error of Astronomical Instruments. 259 



propose to call the reflecting collimator, the error of collimation (or 

 index error as it is generally called) has been read off five times every 

 day, viz. at 6 a, m., at noon, at 6 p. m., at 8 p. m., and at midnight ; 

 taking the mean of these, the error of observation is necessarily very 

 small, and the effect of any accidental difference of temperature in the 

 room, which might alter the figure of the circle at any one time of the 

 day, is at the same time greatly diminished. 



To shew to what extent this mechanical measure, as it may be 

 termed, can be depended upon, I here subjoin the result of the last ten 

 days' observation compared with the index error determined by astro- 

 nomical means, thus : 



Index error of the Madras Mural Circle. 

 By the Reflecting Collimator. By Astronomical Observation. 



No. of Obs. Index Error. NoofObs. Index Error. Difference. 



1835 







/ n 





/ // 





Feb. 



21 



5 



—2-27-36 



6 



—2-27-39 



0-03 





22 



5 



27-92 



8 



27-73 



0-19 





23 



5 



27-64 



9 



27-45 



0-19 





24 



5 



2f>-46 



7 



26-74 



0-28 





25 



5 



27-50 



7 



26-50 



100 





26 



5 



27-22 



6 



27-34 



0-12 





27 



5 



27-28 



9 



2710 



0-18 





28 



5 



26-80 



8 



27-54 



0-74 



March, 



1 



5 



26-91 



9 



27-31 



0-40 





2 



5 



26-83 



9 



27-54 



0-71 



As a further proof of the efficiency of the reflecting collimator, I 

 may adduce the result of observations made at this observatory with 

 the transit instrument. Here we read off twice the sum of the errors 

 of level and collimation, either of which being known leaves us ac- 

 quainted with the other. In the case of the Madras transit instrument, 

 which is furnished with a micrometer, giving motion to a wire parallel 

 to the vertical wires, I have always preferred measuring the error of 

 collimation, and computing the coi-reetions rather than attempting by 

 mechanical adjustment to get rid of it, as is usual with small instruments ; 

 and, on the same principle have always allowed the axis to take up its 

 own position with regard to level ; hence we have only to apply to 

 half the micrometer-reading of the reflecting collimator, the error of 

 level with the proper sign, and the sum or difference, as the case may 

 be, gives the error of collimation, thus : 







Reflecting 



Spirit 



Error of 











Collimation, 



Level, 



Collimation 



Ditto 









or 



or 



by 



by 









L+C 



L 



Refn. Coll. 



Inversion. 



Difference 



1835. 





// 



// 



// 



// 



// 



Feb. 



7 



4*34 



2-60 



1-74 



1-55 



0-19 





9 



4-22 



2-97 



1-25 



1-20 



0-05 





11 



2-92 



2-74 



0-18 



1-00 



1-18 





11 



9-80 



2-74 



*7-06 



5-58 



1-48 





12 



9-28 



2-41 



6-87 



6-73 



0-14 





13 



9-18 



2-61 



6-58 



6'62 



0-04 





16 



9-83 



2-42 



7-41 



6-29 



1-12 





17 



9-97 



1-63 



7-34 



7-17 



0-17 





18 



9-89 



2-04 



7-85 



7-20 



65 





20 



9-37 



2-70 



6-67 



7-19 



0-52 



* I increased the collimation error. 



