INSTRUMENTS.

44. The principal Instruments ordinarily used on Railroad Location are: The Transit, Compass, Level, and Hand Level; and we will consider them in the order here given. (For Instruments used on exploratory-work, see Secs. 141 to 158.)

THE TRANSIT.

Before proceeding with the adjustments of the Transit, it should be seen that the object-glass is screwed firmly home, and a short scratch made on the ring of the glass and continued on to the slide, so that, should the glass be taken out or work loose, it may be screwed up to exactly the same position it was in before. If this is not done, and the glass happens to be badly centred,-i.e., its optical axis does not lie in the centre of the telescope-tube,-if by any chance the glass is moved, the Line of Collimation will also be thrown out of adjustment. The following are the usual adjustments for a Transit :

A. To make the vertical axis truly vertical by means of the small bubble-tubes. Turn the vernier-plate until each of the tubes is parallel to a pair of opposite plate-screws. Bring both bubbles to the centres of the tubes. Then turn the instrument through about 180°. If the bubbles are still in the centre, the adjustment of the small tubes is correct; but if not, correct for half the error in each case by means of the adjusting screws at the ends of the tubes. This adjustment should then be correct; if not, repeat the process until it is.

B. To set the cross-hairs truly vertical and horizontal.--After levelling up, test the vertical hair along its whole length on some fixed point, and if not correct, loosen the capstan-headed screws and move the diaphragm around. The horizontal hair may be tested in a similar way.

C. To make the horizontal axis of the telescope truly horizontal.-Level up the instrument and point the telescope to some object C, as in Fig. 5, at an altitude, if possible, of not less than 45°. Mark the point A where this vertical plane strikes the ground. "Reverse" the instrument, and

if on pointing to C and then reducing to the ground we again strike A, this adjustment is correct.

But suppose the first time the "vertical" plane had struck the ground at B, and then on reversing, instead of striking Bagain, it cuts through some point D. Mark a point E between D and B, distant from D by one quarter of DB. Then by means of the screws under one of the pivots of the horizontal axis bring the intersection of the cross-hairs to strike the point E. This adjustment should then be correct.

D/E A

FIG. 5.

B

D. To make the line of collimation perpendicular to the horizontal axis.-Having levelled up the instrument at O, in Fig. 6, point the telescope to some object C. Turn the telescope over and mark the point A, at a distance 40

equal to about OC, where it strikes the ground in the opposite direction. By making A0 = OC we then obtain a correct adjustment for the line of collimation, even though the objectslide is defective; that is the only reason for making AO and OC about the same length. Reverse, and again point to C; if on turning the telescope over once more it again strikes A, this adjustment is correct. But if instead of intersecting A it cuts through some other point D, then mark a point E between D and B, distant from D by one quarter of DB, and by means of the capstan-headed screws move the diaphragm so as to bring the intersection of the cross-hairs to coincide with E. This adjustment should then be correct. This is liable to throw out adjustment B slightly, so watch that at the same time.

E. To make the long bubble-tube parallel to the line of collimation.-Level up the instrument and clamp the vertical arc. By means of the tangent-screw of the vertical arc bring the bubble to the centre of the tube. Then if the

small bubble-tubes were sufficiently sensitive to render the vertical axis, when the instrument is levelled up, truly vertical, all points cut by the line of collimation equally distant from the instrument would have the same elevation. But it is more satisfactory to obtain a truly vertical axis by means of the long bubble-tube itself, on account of its greater sensitiveness; thus: Level up as accurately as possible by the small tubes, and then treat the long bubble tube as if it were one of the smaller tubes, putting it into a temporary state of adjustment A, by means, not of the screws at the ends of the bubbletube, but by aid of the tangent-screw of the vertical arc, and then by its means obtain a truly vertical axis. Then take the readings on two points A and B equally distant from the instrument and in opposite directions; next move the transit to a point about in the same straight line as A and B, but at as short a distance beyond either of them as the instrument can be focussed to read and level up by the small tubes. Take the reading at A, say 3.43; then if B were previously found to be 1.84 feet higher than A, the telescope should read 1.59 on B if this adjustment were correct. If we do not read this, the screws at the end of the long bubble-tube must be so altered as to bring the bubble to the centre when the instrument reads 1.59. On again pointing to A, the difference between A and B should then be almost 1.59. If it is not quite 1.59, proceed as before until the adjustment is correct.

By moving the instrument into the same line as A and B, as above, we avoid the necessity of levelling up this vertical axis again by means of the long bubble-tube.

Besides the above adjustments, some instruments have a means of Centring the Eye-piece and also of Adjusting the Object-Slide. (See Note C, Appendix.)

45. Remarks.-Another way of performing adjustment C is by means of an object and its reflection in still water, or even in a plate of syrup. A star at night does well for this, but it is advisable to select one as nearly east or west as possible, as its motion in azimuth is then a minimum.

If at any time adjustment C is not correct, we can obtain true results by “reversing,” as in Fig. 5, and remembering that half-way between the two points so found is the correct point.

This latter remark applies also to adjustment D. It is a good plan to reverse on a back-sight every few sights, as it

takes practically no extra time and at once detects if anything is wrong. By taking a point half-way between two points, as D and B in Fig. 6, we can do good work with an instrument in which this adjustment is very far from correct.

As regards adjustment E:-If we had a level handy, it is much more convenient to level two points with it; or if there is a sheet of still water at hand, two pegs driven down to its surface do equally well. To ascertain the Index-error of the vertical circle in instruments where it cannot be corrected for instrumentally, set the vertical axis truly vertical, as explained under adjustment E, then level up the telescope and observe the readings on the vertical arc. If they are at zero, there is no index-error; but if not, the difference between the readings and zero is the index-error.

If the transit has a Striding-level attached, adjustment C may then be more accurately performed by means of it-whether the striding-level is in adjustment itself or not, for it is only the difference of the readings that is required. To make adjustment C then proceed thus: Level up by the small bubbletubes and point the telescope towards the north; take the readings of the bubble on the glass, both at its east and west end; then reverse the striding-level, end for end, and take the readings a second time: one quarter of the difference between the sum of the two east readings and the sum of the two west readings equals the number of divisions on the tube that the bubble must be moved by means of the pivot-screws in order to make the "horizontal axis" level, that end being too high the sum of whose readings is the greater. If the striding-level iş in adjustment, we have only to screw up the "horizontal axis" so as to agree with it. We can, of course, adjust the striding-level by placing it on the pivots already levelled, and bringing the bubble to the centre of the tube.

Lighting the cross-hairs, when the instrument has no lantern attached, can be effected by fastening a piece of bright tin-or even white paper--over and partly in front of the object-glass, so as to cast the reflection of a light on the ground into the tube of the telescope; but the reflector must not obstruct more than half of the field of the object-glass. A piece of tin or paper with a inch hole in the centre of it, fastened at a suitable angle over the object-glass, answers very well.

In moving the diaphragm when the telescope has an invert

ing eye-piece, it has to go in the opposite direction to what appears to be the right one.

If working with an instrument the graduation of which is faulty, read each angle in different parts of the circle. The graduations can always be tested by reading with both verniers on various parts of the circle. In observing an angle, if we take the mean result obtained by both verniers, we eliminate errors due to eccentricity of the vertical axis and the graduated circle, as well as reduce the errors of graduation.

When great accuracy is required in reading an angle the best method to use is BORDA'S REPETITION, which slightly reduces the errors of observation, while it diminishes those of graduation in inverse order to the number of times the angle is repeated. The process is thus: Clamp the vernier-plate to zero, and read the angle by both verniers according to the usual method. Then, keeping the vernier-plate clamped, point the telescope again to the first object, and proceed as before through any number of repetitions. At the end of the final angle read the verniers, adding 360° for each complete révolution which has been made, and divide the total angular measurement by the number of times the angle was repeated. The quotient is the required angle. In this way, provided there is no play about the tangent-screws, an angle can be read with confidence to a few seconds by a very inferior instrument.

In ordinary work, if sure of the correct centring of the vertical axis and also of the graduation itself, there is no need to read by both verniers; but it is advisable to read always by the same vernier if only one is used.

An instrument correct according to the adjustments given above gives correct results when dealing with objects distant from it by the amount OC in Fig. 6, but if there is defective centring of the object-slide-not to be confounded with eccentricity of the optical axis of the object-glass-it will not give correct results in dealing with objects at distances from it greater or less than OC. This can always be tested by ranging points in a 'straight" line for a thousand feet or so, beginning as near to the instrument as the focus will permit. Then, if, on ranging the same points in again from the other end, they do not coincide, one half the difference between the points is the error in alignment. In this way, even with a bad instrument, a straight line can be run. We can of course also run a straight