440 Memorandum on the Construction [No. 138. 



dicular to the brass semicircle A B C, so that the round hole L shall be 

 immediately over the centre point E. 1 1 1 are elevating screws by means 

 of which the instrument is first to be accurately levelled. This may be 

 done either by placing a common spirit level on the brass plate, or by ha- 

 ving two small spirit levels at right angles to each other let into the plate. 

 5 th. To find the meridian line, place the instrument or rather the 

 line E C due north and south by any compass, C being the north 

 point and E the south, at any time (say an hour) before noon* the sun 

 will be observed to shine through the hole L, throwing a bright round spot 

 on the left side of the plate near one of the circles a,f a, a; wait till the 

 spot comes on the circle, say at G, and mark the point with a pencil. Then 

 move the right side of the radius E F up to it, and read off the number 

 of degrees, say 50° on the rim of the plate. The sun after crossing the 

 circle at G will proceed along the dotted line until it reaches the other 

 side of the circle at H, where it will arrive about an hour after noon ; 

 mark the point as before and read off the number of degrees, say 30°, 

 add them to the number noted above (50°), the result will be 80°. 

 Divide by 2, which gives 40°, or the bisection of the arch G H, move 

 the radius to the point of the rim marked 10°, which is half way (or 

 40°J) between the extreme points G and H. The direction of the 

 radius as now placed will be that of the true meridian, being the bi- 

 section of the arch G H, described by the sun himself, (the great 

 Archimedes) at equal altitudes. § The instrument having been origin- 

 ally set to the magnetic meridian, the distance between the line E C 

 and the radius E F; viz. 10°, is the variation east of the compass with 

 which the instrument was set. It is evident that the variation of any 

 other compass may be ascertained in the same manner, or by placing the 

 needle with its own graduated circle on the meridian line E F.|| 



* Two hours would be better if the sun is very high. 



f Any number of circles may be drawn, six or eight are necessary to suit different 

 times of the year. 



\ Either the line M N, or the right side of the radius will answer if placed opposite 

 10°, but the former is best. 



§ The difference in the sun's declination in two or even four hours is so slight, that 

 it would not cause an error in the position of the meridian of more than a few seconds, 

 it is unnecessary therefore to apply the equation of equal altitudes. 



|| Where great accuracy is required, and the needles are long, the observation should 

 be made either at 10 A. m. or 10 p. m. as the needle moves slowly west in the fore- 

 noon, returns to its mean position about 10 p. m. then deviates to the east, and returns 

 at 10 a. m. like the barometer. 



