28 PHILOSOPHICAL TRANSACTIONS. [aNNO 1724. 



the ends were raised so much above the bottom of the box, as to have their 

 upper surfaces, on which the divisions were cut, lie in the same plane with the 

 needle, and at such a distance from each other, that the needle might play freely 

 between them. A few of tlie degrees at the north end were divided into six 

 equal parts, each division being ]0'. It was easy, by the help of a convex glass, 

 to determine the pointing of the needle to less than a quarter of these divisions, 

 or to about 2' of a degree. The pin on which the needle moved, was of steel 

 hardened, and ground to a fine point ; and by a spring placed in the box, the 

 needle might be raised from off the point, and let down again at pleasure, with- 

 out removing the glass, or disturbing the box. By this means both the sharp- 

 ness of the point, and polish of the cap were better preserved from injury, 

 when there was occasion to move the box. A small piece of brass was made to 

 slide on that end of the needle which pointed to the south, for readily bring- 

 ing it to an horizontal position ; for according to the different strength of the 

 touch, the north end of the needle will dip more or less. 



For determining the quantity of the variation, a meridian line was stretched 

 on the top of the house, between the rails of the leads, which were above 15 

 feet asunder. The compass-box was placed on a wooden stool, with 3 feet, that 

 had nothing of iron about it, and its top set level by a plumb-rule. The first 

 needle was a little above -^ of an inch broad, about .o6 in thickness, and 

 weighed about an ounce Troy ; the cap of crystal. After some trials with this 

 needle, it was made narrower, not to exceed half a tenth of an inch, and it 

 then weighed 5 dwt. 5 grains. The 2d needle was at first about -j-V of an inch 

 broad, and .04 thick, the cap of glass ; and after several trials, it was made so 

 much narrower, that its breadth was a little less than its thickness, and it 

 weighed 2 dwt. and 5 grains. The 3d was nearly of tiie same dimensions with 

 the 2d, and weighed 2 dwt. 3 grains. The reason for making the first two 

 needles so heavy, was to try whether they would return more constantly to the 

 same situation than lighter ones. But notwithstanding each of them would 

 settle very exactly in the same place, for a great number of trials made im- 

 mediately one after another, they would differ at times considerably from their 

 former directions. 



This occasioned the making them narrower, fearing their breadth had been 

 some way concerned in this irregularity. But after the alteration, the same 

 thing happened, though nothing of it seemed to proceed from any friction on 

 the point. This made the lighter needles to be preferred, as less apt to injure 

 the point they moved on, and as exact in returning to the same situation. 

 After many trials, Mr. G. found all the needles would not only vary in their 

 direction on difl^erent days, but frequently at different times of the same day ; 



