XX Inteoduction to the Makeestoun Obseevations, 1843. 



June 22*^ 9^ 5™. About the termination of the term observations it was dis- 

 covered that three fibres of the suspension thread were broken ; when the box covers 

 were removed, the stirrup of the magnet was found resting on the copper ring. The 

 observations for some time before this was noticed were found valueless ; and it is 

 probable that they were affected by a varying torsion force throughout the day. 



June 22<i 22^^. While taking out the magnet and suspending the brass bar, the 

 suspension thread broke fibre by fibre. A new suspension thread was prepared a 

 week ago in the following manner : — A quantity of the compound silk fibre, about 

 23 times the length of the desired suspension thread, was run several times from 

 one extremity to the other, between the thumb and index, until almost all the twist 

 which the fibres receive in reeling was removed;* the fibre was then wound side 

 by side round two pins, placed at the required distance in such a manner that no 

 twist was introduced in the winding ; a weight was then suspended by the lower ex- 

 tremity of the loop thus formed, so as to allow the fibres to take the same length, 

 the torsion was so small that the weight did not revolve above 180°. After hang- 

 ing thus for a week, the thread was to-day placed in the declinometer, great care 

 being taken in pegging the extremity, and in winding up to keep the fibres in their 

 respective positions ; this thread was not tied throughout its length, as in the pre- 

 vious case several of the breaks were observed to occur at one of the lower ties. 



June 29. It was found that several trials gave different positions for the plane 

 of detorsion. 



June 29"^ 23^. The plane of detorsion was found to have varied — 74° ; this 

 was eliminated. The brass bar was then lifted two or three times, so as to loosen 

 the fibres, after which the error of the plane of detorsion was found — 26° ; this 

 was also eliminated, and the bar again lifted, when another trial gave the error of the 

 plane of detorsion — 20|^°. After elimination the magnet was inserted. 30^ 3''. The 

 magnet being removed, and the brass bar inserted, the error of the plane of detor- 

 sion was found + 54^°, which was eliminated. It was quite evident that the fibres 

 took diff'erent relative positions too easily, so that, in exchanging the brass bar for 

 the magnet, or vice versa^ considerable torsion might be introduced. At 30** V*, the 

 thread was removed from the declinometer, and suspended on a pin, with the brass 

 bar attached. When the latter came to rest, the thread was tied firmly (but not 

 too tightly) at half-a-dozen places throughout its length with pieces of cotton thread ; 

 the suspension thread was then replaced in the instrument, and, after careful wind- 

 ing up, the torsion was eliminated. This thread, composed of 22 fibres, has an- 

 swered well, improving as it got older. The greater changes of the plane of detor- 

 sion will be found generally connected with some cause stated in the notes to the 

 Daily Observations of Magnetometers. 



18. Accidental sources of error. 



* The fibre is termed untwisted silk ; it is not, however, free of twist, as a slight examination will shew. 



