RECOMMENDATIONS. 487 



able for observations of these faint and often fluctuating me- 

 teors than others of a more refined construction. 



2.) It is recommended that a magnetic needle be kept in a 

 proper place, suspended by a silk fibre or slender hair, (a 

 point-support not being delicate enough,) and so mounted that 

 deviations can be observed to the accuracy of 1'. It has been 

 found convenient to fix in a garden a stone pedestal, on which, 

 at three invariable points, the frame of the magnetic needle 

 rests under a glass cover. The needle, 9 inches long, and of 

 such a weight as to perform about 10 vibrations in a minute, is 

 suspended by one slender hair. There are simple contrivances 

 to steady the needle when required, and to adjust the length 

 of the suspending hair. The scale is divided in degrees for 

 30° on each side of the centre, and in 10' for 1° on each side. 

 There is no vernier, but the place of the needle on the scale 

 is read oft" with great ease by looking through a fixed magni- 

 fying glass, from an opening at some height above, so as to 

 avoid sensible parallax. Professor Christie has described more 

 complete apparatus for this purpose, in the Journal of the 

 Royal Institution, New Series, vol. ii. p. 278. The observer 

 must leave his watch with the assistant, very carefully remove 

 all keys, knives, and other things containing iron, from his 

 dress, and all loose iron tools and utensils, to at least 20 feet 

 distance from the needle. If these precautions are not scru- 

 pulously attended to, the results will be fallacious. It is proper 

 to caution the observer that there is a regular daily variation 

 of the needle, independent of the Aurora. 



Dipping-needles, unless constructed with the utmost care, 

 cannot be considered vei'y satisfactory instruments ; yet, if 

 their suspension be sufficiently delicate, they may probably 

 very well answer for observations during Aurora, of which the 

 object is to determine not the absolute dip of the needle, but 

 the change of dip occasioned by the Aurora. The same pre- 

 cautions of one certain position, removal of iron, &c., are ne- 

 cessary, as in the use of the horizontal needle. 



3.) It is recommended that arrangements be made for ascer- 

 taining the error of a watch. If near an observatory of any 

 kind, the watch should be compared with the transit clock 

 there, immediately after an Aurora : if there is a good meridian 

 line or good dial, the error of the watch on mean time should 

 be found as soon as possible. If a watchmaker in the neigh- 

 bourhood has a good regulator, the watch should be adjusted 

 by it, and the mode of keeping the regulator should be ascer- 

 tained : if a mail-coach from London passes near, the guard's 

 watch may be consulted. The longitude of the place of ob- 



