no iiEroRT— 1881. 



On July 'J15, and again on the 27tL, we tried a series ot' observations 

 with the disturber, in order to determine the absolute value of the scale. 



The guide weight being at a known altitude in the copper tube we took 

 a series of six readings at intervals of a minute, and then shifting the 

 guide weight to another known altitude, took six more in a similar 

 manner ; and so on backwards and forwards for an hour. 



The first movement of the guide weight produced a considerable dis- 

 tu-rbance of an irregular character, and the first set of readings were 

 rejected. Afterwards there was more or less concordance between the 

 results, but it was to be noticed there was a systematic difierence 

 between the change from ' up ' to ' down ' and ' down ' to ' up.' This 

 may perhaps be attributed to friction between certain parts of the 

 apparatus. We believe that on another occasion we might erect the dis- 

 turber under much more favourable conditions, but we do not feel sure 

 that it could ever be made to operate very satisfactorily. 



The series of readings before and after the change of the guide weight 

 were taken in order to determine the path of the pendulum at the critical 

 moment ; but the behaviour of the pendulum is often so irregular, even 

 within a few minutes, that the discrepancy between the several results 

 and the apparent systematic error may be largely due to unknown 

 changes, which took place during the minute which necessarily elapsed 

 between the last of one set of readings and the first of the next. The 

 image took up its new position deliberately, and it was necessary to wait 

 until it had come to its normal position. 



Between the first and second set of observations with the disturber, it 

 had been necessary to enter the room and to recentralise the image. We 

 do not know whether something may not have disturbed the degree of 

 sensitiveness, but at any rate the results of the two sets of observations 

 are very discordant.' 



The first set showed that one inch of movement of the gas-flame, 

 which formed the scale, corresponds with -i^th. of a second of arc of 

 change of horizon ; the second gave j^th of a second to the inch. 



As we can see a twentieth of an inch in the scale, it follows that a 

 change of horizon of about 0"'00.5 should be distinctly visible. In this 

 case the point of the pendulum moves through ^^^ J-y ^th of a millimeter. 

 At present we do not think that the disturber gives more than the order 

 of the changes of horizon which wo note, but our estimate receives a 

 general confirmation frona another circumstance. 



From the delicacy of the gearing connected with the back-leg, we 

 estimate that it is by no means difficult to raise the back-leg by a 

 millionth of an inch. The looseness in the gearing was purposely kept 

 so great that it requires a turn or two of the external pulley on the 

 window-sill before the backlash is absorbed, but after this a very small 

 fraction of a turn is sufficient to move the image in the field. 



We are now inclined to look to this pi'ocess with the back-leg to 

 enable us to determine the actual value of our scale, but this will require 

 a certain amount of new apparatus, which we have not yet had time to 

 arrange. In erecting the instrument we omitted to take certain measure- 

 ments which it now appears will be necessary for the use of the back-leg 

 as a means of determining the absolute value of our scale, but we know 

 these measurements appi'oximately from the working drawings of the 



' See, however, the postscript at the end of this part. 



