275. In the figures 115, 116, the different parts described above are lettered as below. Different makers, however, arrange the parts differently. A is the principal plate, which bears all the other parts. B is the compass-box, sometimes movable about its centre by means of a pinion connected with the milled head I, and capable of being clamped in any position by the screw K. D is the needle, resting on a pivot in the middle of the compass-box. The needle can be raised from its pivot by the screw F. C and C' are the sights, which are fastened to the plate by the screws N. M, M are the levels.

276. The Pivot. This should, as remarked above, be extremely hard and very sharp. It should likewise be placed exactly in the centre of the box and in the line joining the slits in the sights.

To discover whether it is properly centred, and likewise whether the needle is straight, turn the compass until the north point of the needle coincides with any given number of degrees. The south point must be 180° distant. If it is so in all positions, or, in four, distant 90°, as for instance the 0's and 90's, the needle is straight and well centred.

Draw a hair or fine silk string through the slits in the sights. If this passes over the zero-points, the centre is in line. Or, sight to a very near object, and note the reading. Turn the instrument half round, and again note the reading: if these do not agree, the pivot is not on the line of sight. Half the difference is the actual error.

277. The Divided Circle. The accuracy of the division may be tested by turning the plate into different positions. If in all cases the opposite ends of the needle point to the same number of degrees, the probability is that the circle is correctly divided.

If the compass has a vernier, set the instrument in any direction. Then move the box through any number of degrees, and see whether the needle traverses the same number of degrees as the vernier. If it does in all positions, the arc is properly divided.

278. Adjustments. The levels may be adjusted as directed for the transit and theodolite.

The sights should be perpendicular to the plane of the instrument. To verify this, suspend a long plumb-line: level the plate, and sight to this line. If it appears equally distinct through all parts of the slit, the sight is perpenTurn the instrument half round and test the other If either is found incorrect, the

dicular.

sight in the same manner.

maker should rectify it.

279. The compass, as already remarked, is very generally used for surveying purposes, though it is fast giving place to the transit. The latter is furnished with a compass-box, which was not described with the instrument, as it was not needed at that stage of the work. It is in all respects similar to the box attached to the compass itself. The theodolite likewise has a compass. It is, however, so small as to be of very little use in accurate work.

280. The compass is generally supported on an axis inserted in the socket O. This axis terminates in a ball, which works freely but firmly in a socket. This arrangement admits of the axis being placed in any direction. The compass-plate may thus be made level.

Instead of a tripod, many surveyors prefer a single staff pointed with iron. This is called a "Jacob's Staff." Its chief defects are the difficulty of setting in hard ground or among stones, and the want of steadiness in windy weather.

281. Defects of the Compass. Though a very convenient and useful instrument, the compass is deficient in two very important particulars:-its indications are neither correct nor precise.

It is not correct, because, as already remarked, the needle (which is the standard) does not do what it professes: it does not point to the north. This would be of comparatively little importance if its direction were fixed or parallel; but neither of these is the fact. It not only varies

from year to year, but from season to season, and even during the same day. These variations will be the subject of a future chapter.

The presence of ferruginous matter in the earth, or the too great proximity of the chain, or of any other piece of iron, may deflect it very seriously from its normal position.

It is not precise. The divisions on the arc are rarely smaller than half-degrees; and if they were finer it would be difficult to read to less than a quarter of a degree. A little calculation will convince one that this is a serious defect where accuracy is desired. An error of 5' in the bearing would cause a deviation of nearly one foot in ten chains, or about seven feet eight inches in a mile.

SECTION II.

FIELD OPERATIONS.

282. Bearings. To take the bearing of a line, set the compass directly over one end; level it, and turn the plate till the other end of the line-or a rod set up in the direction of the line at a distance as great as is consistent with distinct vision-can be seen through the slits. Then, when the needle has settled, notice the number of degrees to which the end of the needle points, and the cardinal points between which it is situated: the result will be the bearing of the line.

If the north end of the compass is ahead, the north end of the needle should be used, and vice versâ.

If you are running with the north end of the compass ahead, and the north point of the needle is between S. and E. and points to 451°, the bearing is S. 451° E.

In reading, the eye should be placed opposite to the other

end of the needle; otherwise, owing to the parallax of the point, it will appear to stand at a different point of the arc from what it really does. Any iron about the person will be less likely to affect the needle than when in another position.

283. Use of the Vernier. When the needle does not point to one of the divisions of the arc, it is usual to estimate the fraction. Some surveyors, however, after the needle has come to rest, notice between which divisions the needle points, and then move the compass-box, by turning the milled head I, until the point of the needle is opposite one of the divisions. The amount by which the box is turned, as indicated by the vernier, will give the fraction.

This plan, though theoretically correct, adds really nothing to the correctness of the work. The liability to derangement, from handling the instrument, is so great as to neutralize any advantage it might otherwise possess.

284. Reverse Bearing. The reverse bearing of every line should be taken. To do this, set the compass at the position of the rod, and sight back to the former station. The bearing found should be the reverse of the former. If it is not, the work at the former station should be reviewed; if found correct, the difference between the two must arise from some local cause.

285. Local Attraction. When the back sight does not agree with the forward sight, some cause of derangement exists about one of the stations. This is called local attraction. It is generally caused by ferruginous matter in the earth. It is said that any high object, such as a building or even a tree, will slightly deflect the needle. In situations in which trap rocks abound, the local attraction is often very great. The author has known a variation of more than 10° in a line of two and a half chains long, produced by this cause alone. In such regions, running by the needle is very troublesome, and may cause

very serious errors unless great care is taken to allow for the effect produced.

To discover where the attraction exists, select a number of positions in the neighborhood of the suspected points, and note their bearings from these stations, and also from each other. The agreement of several of these will prove their probable correctness. The points thus found to be void of local attraction may be taken as the starting points.

In surveying a farm, a very good way is to note the forward and back sights of every line. If these are found. to agree on any line, they may be presumed to be right, and the others corrected accordingly.

286. To correct for back sights.

When the back sight is greater than the fore sight, subtract the difference from the next bearing, if the two lie between the same points of the compass or between points directly opposite, but add it in all other cases. If the back sight is the less, add the difference in the former case, and subtract it in the latter.

Where the local attraction is great, or the line runs nearly in the direction of one of the cardinal points, a difficulty may occur in the application of the preceding rule. A little reflection will enable the surveyor to modify it to suit the case.

287. By the Vernier. It is more convenient in practice to turn the box by the vernier until the reading for the back sight corresponds with the fore sight. The needle will then give the true bearing of the new line as though no attraction existed.

288. To survey a Farm. Commence by going round it, and verifying, so far as can be done, the landmarks, fixing stakes at the corners, so that the assistant may readily find them if he is not already familiar with their position. Then, placing the compass at one corner,