Page:EB1922 - Volume 32.djvu/653

Set packed for transport.

Receiver and Loop set up ready for use. FIG. 6. Portable Receiving Set.

Receiving Signals.

by the Carl Zeiss optical works of Jena, may be taken. Levels of this kind are also manufactured by Messrs. T. Cooke & Sons of York and Messrs. E. R. Watts & Son of London. The instrument jncludes the following modern improvements which distinguish it from the levels used for rougher surveys or engineering work; the bubble is read from the eye end of the telescope, so that it is not necessary to disturb the instrument by walking to its side and bend- ing over to read it ; the final levelling of the instrument is effected by means of a slow-motion screw which slightly tilts the telescope; and the level is provided with a micrometer attachment in the form of a plate of glass with parallel faces in front of the object glass and this attachment can be tilted by means of a screw, the amount of the tilt being read on a graduated drum. The use of this microm- eter attachment enables readings to be made by the intersection of graduations on the staff, instead of by estimation. It is essential in using this attachment that the sights should be of equal length. Various patterns of staves have been used by different survey de- partments. The Ordnance Survey, for the new geodetic levelling of England and Wales, used staves manufactured by the Cambridge Mathematical Instrument Company. These staves are 10 ft. long and the graduations are marked on a strip of invar (an alloy of steel and nickel which has a very low factor of expansion) ; fixed to each Btaff is a circular level, for the staff-holder to keep it vertical.

Ground Marks. The ground marks to which the observations are taken in the field are of various kinds and are known in British surveying as bench marks. The bench marks in use on the Ordnance Survey are: first class or fundamental; second class, which consist of flush brackets made of bronze let into a wall or other vertical surface, and fixed with cement; rivets, let into horizontal surfaces, such as pavements; and, third class, bench marks cut with a chisel in brick or stone walls, of broad-arrow shape.

The fundamental bench marks are new. It is well known that bench marks of the ordinary type are in general unstable; walls or houses are pulled down, or are subject to settlement, bench marks on isolated stones have been known to be moved and set up in new positions, marks on pavements or kerbs are shifted. None of the old types of bench marks is satisfactory as constituting permanent records of height and position. Moreover, in certain localities, such as mining areas and clay hillsides, the ground itself is unstable. It was therefore decided in planning the geodetic levelling of England and Wales that fundamental bench marks would be devised which should be of a very stable nature. These are established at intervals of about 25 m. from each other; the sites are carefully chosen and no bench mark of this kind is placed on loose soil or rock liable to local disturbance. In constructing such a mark a pit is dug through the soil, sub-soil and loose rock, until sound rock, or hard chalk, is found. The bottom and sides of the pit are lined with concrete, and two reference marks are placed in the bottom concrete, one of bronze and one of polished flint; these are covered with removable caps of metal. When the observations have been finished the pit is filled with sand or other suitable dry material. The internal marks described are those which are used for departmental purposes; but, for the public, an external mark is also provided. It is hoped that these marks will last for many hundreds of years and will, in the future, afford valuable information with regard to vertical move- ments of the crust of the earth.

Method of Observation. The system of carrying out the observa- tions in the field is based on the following principles. In order to minimize the effect of systematic error the levelling of any line is carried out once in each direction; the interval of time between suc- cessive levellings should be as short as possible; to minimize the effect of inclination of the line of sight the distances from the instrument to the fore and back staves should be as nearly equal as possible ; to

enable the graduations on the staves to be read easily the length of a space between the two staves is not in precise work to exceed loo yds. ; to reduce the effect of refraction, no reading is allowed even with the lower stadia hairs nearer than 6 in. from the lower end of a staff; observations are not allowed in bad or windy weather; the level is to be shielded from the direct rays of the sun.

It is on the observance of these and other common sense rules that the accuracy of the work will largely depend, and great attention must be paid to details, such as keeping the staves truly vertical or not letting the staff fall heavily on the picket or bench mark.

Errors. The errors to which levelling is subject may be divided into those due to the staff, those due to the level, those due to the staff-holder, those due to the observer, and those due to the state of the ground, atmospheric conditions and unknown causes. The sys- tem of observation above described is directed towards eliminating, as far as possible, all errors not purely personal. Ultimately, when everything is done to evolve a sound system, it is the human element which tells most in the result, and the observer should possess ex- cellent eyesight, a good stock of patience and be scrupulously honest ; for in this, as in all scientific measurement, there must not be the least bias or wish to obtain a particular result. If the errors were alt accidental and subject to the ordinary law, the probable error should increase as the square root of the distance levelled. There is, however, in levelling, a factor known as the " systematic error," by which the far end of a line constantly tends to appear the lower. If this systematic error were quite uniform, then double levelling would completely eliminate it; but this is not quite the case, and in the result, we are left with errors mainly accidental, plus an un- known amount of error not strictly subject to the law of accidental errors. A careful investigation by the Ordnance Survey has shown that the safest course, and that most in accordance with the con- ditions of the case, is to treat the whole error as accidental and as accumulating in proportion to the square root of the length of the line levelled. This is contrary to French practice. The probable error of I m. of double levelling is usually calculated from the formula =0^67 (Sd 2 /4M), where "ZcP is the sum of the squares of the discrepancies between the forward and back levelling from mark to mark, and M the length of the line in miles. The value of the probable error so obtained will usually be less than that obtained from a consideration of the closing errors. In recent Ordnance Sur- vey precise levelling the value of e, from the above formula, is found to be somewhat less than -003 ft., whereas the value found from a discussion of the actual errors of closure_of the level net-work is 0077. The probable error accumulated in the net-work between Newlyn in Cornwall and Dunbar in Haddingtonshire, two places separated by about 700 m. of levelling, is 0-16 ft., or about 2 in. This figure will serve to give an idea of the accuracy of modern pre- cise levelling.

AUTHORITIES. See Close & Cox, Text Book of Topographical and Geographical Surveying i(H.M. Stationery Office, London, 1913); Middleton & Chad wick, A Treatise on Surveying (1911), chiefly of value for engineering surveys; W. Norman Thomas, Surveying (1920). (C. F. CL. ; H. S. L. W.)

NAUTICAL SURVEYING

Although the World War restricted hydrographic surveying work, it led to improvements in methods and in instruments.

So far as the British Naval Surveying Service was concerned, it had generally been considered before that a marine survey need not be carried out with the degree of accuracy which is rightly considered necessary in purely land surveys; it being